diff --git a/source/devices/AM62DX/linux/Linux_Performance_Guide.rst b/source/devices/AM62DX/linux/Linux_Performance_Guide.rst index 51d28dc65..343f71156 100644 --- a/source/devices/AM62DX/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62DX/linux/Linux_Performance_Guide.rst @@ -1,10 +1,11 @@ -==================================== - Linux 11.02.08.02 Performance Guide -==================================== +################################# + Linux 12.00.00 Performance Guide +################################# -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -18,8 +19,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -27,17 +29,20 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ +| + +***************** System Benchmarks ------------------ +***************** + +| LMBench -^^^^^^^ +======= + LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at @@ -64,124 +69,125 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "af_unix_sock_stream_latency (microsec)","30.12 (min 29.31, max 30.82)" - "af_unix_socket_stream_bandwidth (mb\s)","1054.72 (min 1042.80, max 1065.78)" - "bw_file_rd-io-1mb (mb/s)","1394.63 (min 1375.04, max 1403.65)" - "bw_file_rd-o2c-1mb (mb/s)","749.43 (min 694.32, max 788.64)" - "bw_mem-bcopy-16mb (mb/s)","1740.68 (min 1715.63, max 1750.16)" - "bw_mem-bcopy-1mb (mb/s)","1863.03 (min 1842.75, max 1889.39)" - "bw_mem-bcopy-2mb (mb/s)","1625.67 (min 1545.83, max 1692.05)" - "bw_mem-bcopy-4mb (mb/s)","1647.32 (min 1548.79, max 1713.31)" - "bw_mem-bcopy-8mb (mb/s)","1729.40 (min 1707.76, max 1746.34)" - "bw_mem-bzero-16mb (mb/s)","7938.98 (min 7920.79, max 7952.29)" - "bw_mem-bzero-1mb (mb/s)","4881.67 (min 1842.75, max 7925.07)" - "bw_mem-bzero-2mb (mb/s)","4766.95 (min 1545.83, max 7952.29)" - "bw_mem-bzero-4mb (mb/s)","4788.22 (min 1548.79, max 7945.11)" - "bw_mem-bzero-8mb (mb/s)","4831.41 (min 1707.76, max 7947.02)" - "bw_mem-cp-16mb (mb/s)","840.43 (min 834.03, max 852.65)" - "bw_mem-cp-1mb (mb/s)","4506.74 (min 795.42, max 8218.15)" - "bw_mem-cp-2mb (mb/s)","4415.13 (min 784.93, max 8064.52)" - "bw_mem-cp-4mb (mb/s)","4431.16 (min 849.98, max 8001.45)" - "bw_mem-cp-8mb (mb/s)","4434.57 (min 870.70, max 7984.03)" - "bw_mem-fcp-16mb (mb/s)","1532.91 (min 1509.72, max 1557.94)" - "bw_mem-fcp-1mb (mb/s)","4726.56 (min 1515.15, max 7925.07)" - "bw_mem-fcp-2mb (mb/s)","4700.63 (min 1429.08, max 7952.29)" - "bw_mem-fcp-4mb (mb/s)","4735.69 (min 1504.61, max 7945.11)" - "bw_mem-fcp-8mb (mb/s)","4744.51 (min 1547.39, max 7947.02)" - "bw_mem-frd-16mb (mb/s)","1887.64 (min 1876.61, max 1895.06)" - "bw_mem-frd-1mb (mb/s)","1782.58 (min 1515.15, max 2047.08)" - "bw_mem-frd-2mb (mb/s)","1605.52 (min 1429.08, max 1839.25)" - "bw_mem-frd-4mb (mb/s)","1646.92 (min 1504.61, max 1855.86)" - "bw_mem-frd-8mb (mb/s)","1665.51 (min 1547.39, max 1882.13)" - "bw_mem-fwr-16mb (mb/s)","7962.52 (min 7943.07, max 7974.75)" - "bw_mem-fwr-1mb (mb/s)","5099.94 (min 1932.20, max 8218.15)" - "bw_mem-fwr-2mb (mb/s)","4869.22 (min 1611.08, max 8064.52)" - "bw_mem-fwr-4mb (mb/s)","4872.60 (min 1653.58, max 8001.45)" - "bw_mem-fwr-8mb (mb/s)","4875.39 (min 1660.27, max 7984.03)" - "bw_mem-rd-16mb (mb/s)","1932.53 (min 1913.65, max 1942.22)" - "bw_mem-rd-1mb (mb/s)","1725.48 (min 1225.28, max 2219.76)" - "bw_mem-rd-2mb (mb/s)","1571.95 (min 1133.63, max 1943.32)" - "bw_mem-rd-4mb (mb/s)","1665.97 (min 1365.65, max 1941.12)" - "bw_mem-rd-8mb (mb/s)","1730.15 (min 1454.28, max 1948.61)" - "bw_mem-rdwr-16mb (mb/s)","1672.28 (min 1600.32, max 1718.40)" - "bw_mem-rdwr-1mb (mb/s)","1037.11 (min 795.42, max 1297.02)" - "bw_mem-rdwr-2mb (mb/s)","986.71 (min 784.93, max 1238.39)" - "bw_mem-rdwr-4mb (mb/s)","1135.79 (min 849.98, max 1446.39)" - "bw_mem-rdwr-8mb (mb/s)","1211.02 (min 870.70, max 1599.04)" - "bw_mem-wr-16mb (mb/s)","1619.91 (min 1548.29, max 1668.40)" - "bw_mem-wr-1mb (mb/s)","1255.05 (min 1206.06, max 1308.90)" - "bw_mem-wr-2mb (mb/s)","1195.30 (min 1057.64, max 1289.08)" - "bw_mem-wr-4mb (mb/s)","1403.98 (min 1295.97, max 1454.55)" - "bw_mem-wr-8mb (mb/s)","1525.80 (min 1389.37, max 1599.04)" - "bw_mmap_rd-mo-1mb (mb/s)","2117.81 (min 2081.76, max 2139.04)" - "bw_mmap_rd-o2c-1mb (mb/s)","750.19 (min 693.72, max 794.91)" - "bw_pipe (mb/s)","702.59 (min 692.16, max 713.48)" - "bw_unix (mb/s)","1054.72 (min 1042.80, max 1065.78)" - "lat_connect (us)","57.00 (min 56.61, max 57.41)" - "lat_ctx-2-128k (us)","7.85 (min 7.55, max 8.11)" - "lat_ctx-2-256k (us)","7.33 (min 6.40, max 8.28)" - "lat_ctx-4-128k (us)","7.44 (min 7.06, max 7.89)" - "lat_ctx-4-256k (us)","7.12 (min 5.25, max 8.97)" - "lat_fs-0k (num_files)","234.50 (min 215.00, max 251.00)" - "lat_fs-10k (num_files)","113.25 (min 104.00, max 122.00)" - "lat_fs-1k (num_files)","163.75 (min 144.00, max 178.00)" - "lat_fs-4k (num_files)","159.75 (min 144.00, max 174.00)" - "lat_mem_rd-stride128-sz1000k (ns)","31.04 (min 30.78, max 31.41)" - "lat_mem_rd-stride128-sz125k (ns)","5.56 (min 5.53, max 5.61)" - "lat_mem_rd-stride128-sz250k (ns)","5.90 (min 5.83, max 6.36)" - "lat_mem_rd-stride128-sz31k (ns)","3.67 (min 2.16, max 4.19)" + "af_unix_sock_stream_latency (microsec)","26.25 (min 23.23, max 29.86)" + "af_unix_socket_stream_bandwidth (mb\s)","1071.71 (min 1058.15, max 1084.28)" + "bw_file_rd-io-1mb (mb/s)","1348.18 (min 1335.78, max 1359.22)" + "bw_file_rd-o2c-1mb (mb/s)","696.88 (min 633.61, max 771.01)" + "bw_mem-bcopy-16mb (mb/s)","1742.80 (min 1730.67, max 1760.18)" + "bw_mem-bcopy-1mb (mb/s)","1865.40 (min 1833.37, max 1893.94)" + "bw_mem-bcopy-2mb (mb/s)","1625.86 (min 1561.59, max 1675.32)" + "bw_mem-bcopy-4mb (mb/s)","1648.95 (min 1618.78, max 1670.84)" + "bw_mem-bcopy-8mb (mb/s)","1674.23 (min 1590.62, max 1730.10)" + "bw_mem-bzero-16mb (mb/s)","7939.79 (min 7935.20, max 7947.02)" + "bw_mem-bzero-1mb (mb/s)","4882.93 (min 1833.37, max 7950.36)" + "bw_mem-bzero-2mb (mb/s)","4763.62 (min 1561.59, max 7916.52)" + "bw_mem-bzero-4mb (mb/s)","4781.67 (min 1618.78, max 7926.50)" + "bw_mem-bzero-8mb (mb/s)","4802.10 (min 1590.62, max 7943.07)" + "bw_mem-cp-16mb (mb/s)","841.31 (min 833.20, max 851.06)" + "bw_mem-cp-1mb (mb/s)","4496.61 (min 783.70, max 8222.18)" + "bw_mem-cp-2mb (mb/s)","4414.12 (min 780.03, max 8041.96)" + "bw_mem-cp-4mb (mb/s)","4417.08 (min 815.58, max 8002.91)" + "bw_mem-cp-8mb (mb/s)","4429.55 (min 846.20, max 7980.05)" + "bw_mem-fcp-16mb (mb/s)","1554.75 (min 1535.21, max 1576.04)" + "bw_mem-fcp-1mb (mb/s)","4733.08 (min 1539.78, max 7950.36)" + "bw_mem-fcp-2mb (mb/s)","4701.07 (min 1459.06, max 7916.52)" + "bw_mem-fcp-4mb (mb/s)","4726.03 (min 1495.51, max 7926.50)" + "bw_mem-fcp-8mb (mb/s)","4740.18 (min 1540.24, max 7943.07)" + "bw_mem-frd-16mb (mb/s)","1841.28 (min 1782.13, max 1859.38)" + "bw_mem-frd-1mb (mb/s)","1754.60 (min 1539.78, max 1975.57)" + "bw_mem-frd-2mb (mb/s)","1618.86 (min 1459.06, max 1750.09)" + "bw_mem-frd-4mb (mb/s)","1648.13 (min 1495.51, max 1827.32)" + "bw_mem-frd-8mb (mb/s)","1658.91 (min 1540.24, max 1849.92)" + "bw_mem-fwr-16mb (mb/s)","7957.56 (min 7949.65, max 7964.16)" + "bw_mem-fwr-1mb (mb/s)","5064.88 (min 1899.83, max 8222.18)" + "bw_mem-fwr-2mb (mb/s)","4873.22 (min 1723.84, max 8041.96)" + "bw_mem-fwr-4mb (mb/s)","4873.01 (min 1684.21, max 8002.91)" + "bw_mem-fwr-8mb (mb/s)","4869.76 (min 1689.72, max 7980.05)" + "bw_mem-rd-16mb (mb/s)","1930.96 (min 1920.08, max 1941.51)" + "bw_mem-rd-1mb (mb/s)","1704.69 (min 1174.50, max 2232.56)" + "bw_mem-rd-2mb (mb/s)","1532.41 (min 1099.30, max 1964.96)" + "bw_mem-rd-4mb (mb/s)","1624.88 (min 1203.55, max 1922.15)" + "bw_mem-rd-8mb (mb/s)","1714.81 (min 1360.78, max 1945.05)" + "bw_mem-rdwr-16mb (mb/s)","1609.39 (min 1560.52, max 1682.09)" + "bw_mem-rdwr-1mb (mb/s)","1008.69 (min 783.70, max 1286.29)" + "bw_mem-rdwr-2mb (mb/s)","970.31 (min 780.03, max 1168.68)" + "bw_mem-rdwr-4mb (mb/s)","1109.24 (min 815.58, max 1465.74)" + "bw_mem-rdwr-8mb (mb/s)","1224.13 (min 846.20, max 1571.40)" + "bw_mem-wr-16mb (mb/s)","1620.26 (min 1557.03, max 1666.84)" + "bw_mem-wr-1mb (mb/s)","1214.54 (min 1156.07, max 1286.29)" + "bw_mem-wr-2mb (mb/s)","1139.75 (min 1096.49, max 1257.66)" + "bw_mem-wr-4mb (mb/s)","1356.70 (min 1203.55, max 1465.74)" + "bw_mem-wr-8mb (mb/s)","1534.69 (min 1360.78, max 1571.40)" + "bw_mmap_rd-mo-1mb (mb/s)","2054.67 (min 2010.97, max 2106.47)" + "bw_mmap_rd-o2c-1mb (mb/s)","721.65 (min 636.84, max 812.35)" + "bw_pipe (mb/s)","712.68 (min 661.13, max 740.02)" + "bw_unix (mb/s)","1071.71 (min 1058.15, max 1084.28)" + "lat_connect (us)","61.57 (min 50.64, max 72.15)" + "lat_ctx-2-128k (us)","8.53 (min 7.89, max 9.31)" + "lat_ctx-2-256k (us)","18.50 (min 6.99, max 40.68)" + "lat_ctx-4-128k (us)","14.98 (min 8.17, max 29.59)" + "lat_ctx-4-256k (us)","22.67 (min 7.08, max 97.51)" + "lat_fs-0k (num_files)","280.00 (min 256.00, max 311.00)" + "lat_fs-10k (num_files)","123.50 (min 105.00, max 141.00)" + "lat_fs-1k (num_files)","187.17 (min 173.00, max 203.00)" + "lat_fs-4k (num_files)","180.67 (min 163.00, max 207.00)" + "lat_mem_rd-stride128-sz1000k (ns)","30.92 (min 30.56, max 31.60)" + "lat_mem_rd-stride128-sz125k (ns)","5.57 (min 5.54, max 5.63)" + "lat_mem_rd-stride128-sz250k (ns)","5.83 (min 5.82, max 5.84)" + "lat_mem_rd-stride128-sz31k (ns)","3.49 (min 2.15, max 4.20)" "lat_mem_rd-stride128-sz50 (ns)","2.15" - "lat_mem_rd-stride128-sz500k (ns)","11.25 (min 9.86, max 13.27)" - "lat_mem_rd-stride128-sz62k (ns)","5.15 (min 4.51, max 5.27)" - "lat_mmap-1m (us)","55.75 (min 51.00, max 58.00)" - "lat_ops-double-add (ns)","2.86" - "lat_ops-double-div (ns)","15.75 (min 15.74, max 15.76)" + "lat_mem_rd-stride128-sz500k (ns)","11.53 (min 10.19, max 12.71)" + "lat_mem_rd-stride128-sz62k (ns)","5.23 (min 5.19, max 5.25)" + "lat_mmap-1m (us)","53.33 (min 49.00, max 58.00)" + "lat_ops-double-add (ns)","2.86 (min 2.86, max 2.87)" + "lat_ops-double-div (ns)","15.74" "lat_ops-double-mul (ns)","2.86" - "lat_ops-float-add (ns)","2.86" + "lat_ops-float-add (ns)","2.86 (min 2.86, max 2.87)" "lat_ops-float-div (ns)","9.30 (min 9.30, max 9.31)" "lat_ops-float-mul (ns)","2.86" "lat_ops-int-add (ns)","0.72" "lat_ops-int-bit (ns)","0.48" - "lat_ops-int-div (ns)","4.29" + "lat_ops-int-div (ns)","4.30 (min 4.29, max 4.30)" "lat_ops-int-mod (ns)","4.53 (min 4.53, max 4.54)" - "lat_ops-int-mul (ns)","3.08 (min 3.07, max 3.09)" + "lat_ops-int-mul (ns)","3.09 (min 3.04, max 3.12)" "lat_ops-int64-add (ns)","0.72" "lat_ops-int64-bit (ns)","0.48" "lat_ops-int64-div (ns)","6.80" - "lat_ops-int64-mod (ns)","5.25 (min 5.25, max 5.26)" - "lat_ops-int64-mul (ns)","3.55" - "lat_pagefault (us)","0.65 (min 0.51, max 1.06)" - "lat_pipe (us)","25.70 (min 25.35, max 26.22)" - "lat_proc-exec (us)","712.79 (min 689.13, max 735.50)" - "lat_proc-fork (us)","624.24 (min 601.00, max 643.00)" + "lat_ops-int64-mod (ns)","5.25" + "lat_ops-int64-mul (ns)","3.57 (min 3.54, max 3.64)" + "lat_pagefault (us)","0.44" + "lat_pipe (us)","21.45 (min 20.89, max 21.91)" + "lat_proc-exec (us)","714.75 (min 702.50, max 721.88)" + "lat_proc-fork (us)","632.38 (min 613.13, max 654.78)" "lat_proc-proccall (us)","0.01" - "lat_select (us)","34.08 (min 33.89, max 34.33)" - "lat_sem (us)","2.99 (min 2.61, max 3.63)" - "lat_sig-catch (us)","5.51 (min 5.27, max 5.71)" - "lat_sig-install (us)","0.67 (min 0.64, max 0.70)" - "lat_sig-prot (us)","0.64 (min 0.46, max 0.80)" - "lat_syscall-fstat (us)","1.98 (min 1.90, max 2.07)" - "lat_syscall-null (us)","0.46 (min 0.46, max 0.50)" - "lat_syscall-open (us)","166.27 (min 150.75, max 200.85)" - "lat_syscall-read (us)","0.82 (min 0.80, max 0.88)" - "lat_syscall-stat (us)","4.79 (min 4.64, max 4.97)" - "lat_syscall-write (us)","0.78 (min 0.75, max 0.83)" - "lat_tcp (us)","0.92 (min 0.91, max 0.97)" - "lat_unix (us)","30.12 (min 29.31, max 30.82)" - "latency_for_0.50_mb_block_size (nanosec)","11.25 (min 9.86, max 13.27)" - "latency_for_1.00_mb_block_size (nanosec)","15.52 (min 0.00, max 31.41)" - "pipe_bandwidth (mb\s)","702.59 (min 692.16, max 713.48)" - "pipe_latency (microsec)","25.70 (min 25.35, max 26.22)" + "lat_select (us)","31.59 (min 31.34, max 31.76)" + "lat_sem (us)","2.73 (min 2.22, max 3.05)" + "lat_sig-catch (us)","5.60 (min 5.47, max 5.74)" + "lat_sig-install (us)","0.65 (min 0.58, max 0.71)" + "lat_sig-prot (us)","0.85 (min 0.71, max 0.92)" + "lat_syscall-fstat (us)","1.81 (min 1.74, max 1.87)" + "lat_syscall-null (us)","0.44 (min 0.37, max 0.51)" + "lat_syscall-open (us)","206.84 (min 162.68, max 241.21)" + "lat_syscall-read (us)","0.75 (min 0.73, max 0.79)" + "lat_syscall-stat (us)","4.15 (min 3.99, max 4.26)" + "lat_syscall-write (us)","0.70 (min 0.66, max 0.75)" + "lat_tcp (us)","0.89 (min 0.76, max 1.02)" + "lat_unix (us)","26.25 (min 23.23, max 29.86)" + "latency_for_0.50_mb_block_size (nanosec)","11.53 (min 10.19, max 12.71)" + "latency_for_1.00_mb_block_size (nanosec)","15.46 (min 0.00, max 31.60)" + "pipe_bandwidth (mb\s)","712.68 (min 661.13, max 740.02)" + "pipe_latency (microsec)","21.45 (min 20.89, max 21.91)" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","34.08 (min 33.89, max 34.33)" - "semaphore_latency (microsec)","2.99 (min 2.61, max 3.63)" - "signal_handler_latency (microsec)","0.67 (min 0.64, max 0.70)" - "signal_handler_overhead (microsec)","5.51 (min 5.27, max 5.71)" - "tcp_ip_connection_cost_to_localhost (microsec)","57.00 (min 56.61, max 57.41)" - "tcp_latency_using_localhost (microsec)","0.92 (min 0.91, max 0.97)" + "select_on_200_tcp_fds (microsec)","31.59 (min 31.34, max 31.76)" + "semaphore_latency (microsec)","2.73 (min 2.22, max 3.05)" + "signal_handler_latency (microsec)","0.65 (min 0.58, max 0.71)" + "signal_handler_overhead (microsec)","5.60 (min 5.47, max 5.74)" + "tcp_ip_connection_cost_to_localhost (microsec)","61.57 (min 50.64, max 72.15)" + "tcp_latency_using_localhost (microsec)","0.89 (min 0.76, max 1.02)" Dhrystone -^^^^^^^^^ +========= + Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. @@ -199,11 +205,12 @@ Execute the benchmark with the following: :header: "Benchmarks","am62dxx_evm-fs: perf" "cpu_clock (mhz)","1400.00" - "dhrystone_per_mhz (dmips/mhz)","2.90" - "dhrystone_per_second (dhrystonep)","7142857.00" + "dhrystone_per_mhz (dmips/mhz)","2.75 (min 2.60, max 2.90)" + "dhrystone_per_second (dhrystonep)","6789289.58 (min 6451613.00, max 7142857.00)" Whetstone -^^^^^^^^^ +========= + Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -215,20 +222,22 @@ Execute the benchmark with the following: .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "whetstone (mips)","7500.00 (min 5000.00, max 10000.00)" + "whetstone (mips)","5833.33 (min 5000.00, max 10000.00)" Linpack -^^^^^^^ +======= + Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "linpack (kflops)","576688.25 (min 574136.00, max 578099.00)" + "linpack (kflops)","580360.75 (min 579515.00, max 581724.00)" Stream -^^^^^^ +====== + STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -248,13 +257,14 @@ Execute the benchmark with the following: .. csv-table:: Stream Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "add (mb/s)","2565.85 (min 2558.60, max 2572.00)" - "copy (mb/s)","3569.39 (min 3553.00, max 3582.00)" - "scale (mb/s)","3338.13 (min 3282.60, max 3379.00)" - "triad (mb/s)","2385.28 (min 2379.40, max 2389.10)" + "add (mb/s)","2566.47 (min 2557.40, max 2582.90)" + "copy (mb/s)","3540.12 (min 3515.30, max 3565.80)" + "scale (mb/s)","3346.33 (min 3203.90, max 3405.60)" + "triad (mb/s)","2384.27 (min 2377.90, max 2393.60)" CoreMarkPro -^^^^^^^^^^^ +=========== + CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, @@ -264,33 +274,35 @@ and floating-point workloads, and data sets for utilizing larger memory subsyste .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "cjpeg-rose7-preset (workloads/)","42.00 (min 41.84, max 42.19)" + "cjpeg-rose7-preset (workloads/)","41.99 (min 41.84, max 42.02)" "core (workloads/)","0.30" - "coremark-pro ()","923.79 (min 903.66, max 936.49)" - "linear_alg-mid-100x100-sp (workloads/)","14.68 (min 14.67, max 14.69)" + "coremark-pro ()","928.67 (min 892.10, max 957.91)" + "linear_alg-mid-100x100-sp (workloads/)","14.68 (min 14.68, max 14.69)" "loops-all-mid-10k-sp (workloads/)","0.71" - "nnet_test (workloads/)","1.09 (min 1.08, max 1.09)" - "parser-125k (workloads/)","8.78 (min 8.70, max 8.85)" - "radix2-big-64k (workloads/)","61.55 (min 51.68, max 69.15)" - "sha-test (workloads/)","81.30 (min 80.65, max 81.97)" - "zip-test (workloads/)","22.16 (min 21.74, max 22.22)" + "nnet_test (workloads/)","1.08" + "parser-125k (workloads/)","8.96 (min 8.77, max 9.26)" + "radix2-big-64k (workloads/)","63.08 (min 44.67, max 75.93)" + "sha-test (workloads/)","81.19 (min 80.65, max 81.97)" + "zip-test (workloads/)","22.66 (min 22.22, max 23.81)" + .. csv-table:: CoreMarkProFourCore Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "cjpeg-rose7-preset (workloads/)","160.97 (min 158.73, max 161.29)" + "cjpeg-rose7-preset (workloads/)","159.05 (min 158.73, max 161.29)" "core (workloads/)","1.20" - "coremark-pro ()","2540.92 (min 2504.79, max 2570.41)" - "linear_alg-mid-100x100-sp (workloads/)","56.35 (min 56.31, max 56.37)" - "loops-all-mid-10k-sp (workloads/)","2.07 (min 2.03, max 2.12)" - "nnet_test (workloads/)","3.62" - "parser-125k (workloads/)","9.29 (min 8.89, max 9.93)" - "radix2-big-64k (workloads/)","81.13 (min 79.71, max 82.65)" - "sha-test (workloads/)","269.38 (min 263.16, max 270.27)" - "zip-test (workloads/)","75.50 (min 72.73, max 76.92)" + "coremark-pro ()","2592.85 (min 2558.16, max 2639.06)" + "linear_alg-mid-100x100-sp (workloads/)","56.35 (min 56.24, max 56.43)" + "loops-all-mid-10k-sp (workloads/)","2.05 (min 2.01, max 2.11)" + "nnet_test (workloads/)","3.59" + "parser-125k (workloads/)","11.93 (min 9.98, max 13.65)" + "radix2-big-64k (workloads/)","76.45 (min 71.88, max 81.33)" + "sha-test (workloads/)","266.71 (min 263.16, max 270.27)" + "zip-test (workloads/)","78.74 (min 74.07, max 83.33)" MultiBench -^^^^^^^^^^ +========== + MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -309,52 +321,55 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am62dxx_evm-fs: perf" - "4m-check (workloads/)","417.14 (min 415.49, max 418.62)" - "4m-check-reassembly (workloads/)","113.35 (min 111.86, max 115.08)" - "4m-check-reassembly-tcp (workloads/)","58.41 (min 58.14, max 58.69)" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","33.11 (min 32.89, max 33.33)" - "4m-check-reassembly-tcp-x264w2 (workloads/)","1.89 (min 1.87, max 1.91)" - "4m-cmykw2 (workloads/)","240.20 (min 227.02, max 246.61)" - "4m-cmykw2-rotatew2 (workloads/)","49.45 (min 49.14, max 49.75)" - "4m-reassembly (workloads/)","79.96 (min 78.86, max 80.97)" - "4m-rotatew2 (workloads/)","52.44 (min 51.84, max 52.83)" - "4m-tcp-mixed (workloads/)","119.07 (min 118.52, max 120.30)" - "4m-x264w2 (workloads/)","1.97 (min 1.94, max 2.00)" - "idct-4m (workloads/)","19.16 (min 19.14, max 19.20)" - "idct-4mw1 (workloads/)","19.17 (min 19.15, max 19.21)" - "ippktcheck-4m (workloads/)","414.97 (min 413.43, max 417.08)" - "ippktcheck-4mw1 (workloads/)","416.58 (min 413.22, max 418.76)" - "ipres-4m (workloads/)","108.42 (min 107.30, max 109.49)" - "ipres-4mw1 (workloads/)","107.73 (min 106.84, max 109.09)" - "md5-4m (workloads/)","27.55 (min 27.30, max 27.87)" - "md5-4mw1 (workloads/)","27.59 (min 27.32, max 27.78)" - "rgbcmyk-4m (workloads/)","63.65 (min 63.57, max 63.78)" - "rgbcmyk-4mw1 (workloads/)","63.67 (min 63.41, max 63.88)" - "rotate-4ms1 (workloads/)","23.44 (min 23.33, max 23.55)" - "rotate-4ms1w1 (workloads/)","23.47 (min 23.28, max 23.84)" - "rotate-4ms64 (workloads/)","23.71 (min 23.61, max 23.79)" - "rotate-4ms64w1 (workloads/)","23.70 (min 23.57, max 23.76)" + "4m-check (workloads/)","423.43 (min 421.30, max 428.08)" + "4m-check-reassembly (workloads/)","122.51 (min 108.93, max 172.71)" + "4m-check-reassembly-tcp (workloads/)","62.57 (min 59.52, max 73.75)" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","34.10 (min 33.67, max 34.90)" + "4m-check-reassembly-tcp-x264w2 (workloads/)","1.89 (min 1.88, max 1.89)" + "4m-cmykw2 (workloads/)","245.65 (min 239.52, max 250.00)" + "4m-cmykw2-rotatew2 (workloads/)","47.70 (min 46.77, max 49.75)" + "4m-reassembly (workloads/)","91.00 (min 76.45, max 147.06)" + "4m-rotatew2 (workloads/)","52.53 (min 51.79, max 54.11)" + "4m-tcp-mixed (workloads/)","128.00 (min 126.98, max 129.03)" + "4m-x264w2 (workloads/)","1.95 (min 1.93, max 1.98)" + "idct-4m (workloads/)","19.28 (min 19.16, max 19.76)" + "idct-4mw1 (workloads/)","19.28 (min 19.15, max 19.74)" + "ippktcheck-4m (workloads/)","422.45 (min 418.62, max 428.52)" + "ippktcheck-4mw1 (workloads/)","422.78 (min 419.96, max 426.40)" + "ipres-4m (workloads/)","101.92 (min 99.60, max 107.84)" + "ipres-4mw1 (workloads/)","102.01 (min 99.87, max 108.38)" + "md5-4m (workloads/)","28.00 (min 27.77, max 28.14)" + "md5-4mw1 (workloads/)","28.03 (min 27.97, max 28.12)" + "rgbcmyk-4m (workloads/)","63.80 (min 63.41, max 64.54)" + "rgbcmyk-4mw1 (workloads/)","63.80 (min 63.45, max 64.45)" + "rotate-4ms1 (workloads/)","23.83 (min 23.29, max 25.80)" + "rotate-4ms1w1 (workloads/)","23.84 (min 23.28, max 25.80)" + "rotate-4ms64 (workloads/)","24.14 (min 23.56, max 26.11)" + "rotate-4ms64w1 (workloads/)","24.14 (min 23.57, max 26.08)" "x264-4mq (workloads/)","0.58" "x264-4mqw1 (workloads/)","0.58" +| + Boot-time Measurement ---------------------- +===================== Boot media: MMCSD -^^^^^^^^^^^^^^^^^ +----------------- .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am62dxx_evm-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","14.31 (min 12.88, max 20.51)" + "Linux boot time from SD with default rootfs (20 boot cycles)","17.16 (min 16.40, max 18.33)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. | Ethernet ------------------ -Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html +======== + +Ethernet performance benchmarks were measured using :command:`netperf` 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, @@ -364,11 +379,11 @@ CPU utilization was captured as the total percentage used across all cores on th while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. -In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth +In this scenario, :command:`netperf` options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: -:: +.. code-block:: console burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) burst_size = 500000000 / 8 / 1472 / 100 = 425 @@ -376,21 +391,21 @@ no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B data wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the :command:`netperf` test with no bandwidth limit (remove -b/-w options). -In order to start a netperf client on one device, the other device must have netserver running. -To start netserver: +In order to start a :command:`netperf` client on one device, the other device must have :command:`netserver` running. +To start :command:`netserver`: -:: +.. code-block:: console netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger :command:`netperf` clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test. -:: +.. code-block:: console #!/bin/bash for i in 1 @@ -402,179 +417,142 @@ additional timing measurements during the test. -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for +Running the following commands will trigger :command:`netperf` clients to measure UDP burst performance for 60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run :command:`netperf` client from DUT and start :command:`netserver` on tester. -:: +.. code-block:: console - netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m - -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m + -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run :command:`netperf` client from tester and start :command:`netserver` on DUT. -:: +.. code-block:: console netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE +CPSW/CPSW2g/CPSW3g Ethernet +--------------------------- -CPSW/CPSW2g/CPSW3g Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -- CPSW2g: AM65x, J7200, J721e, J721S2, J784S4, J742S2 -- CPSW3g: AM64x, AM62x, AM62ax, AM62px, AM62dx - - -.. rubric:: TCP Bidirectional Throughput - :name: CPSW2g-tcp-bidirectional-throughput - -.. csv-table:: CPSW2g TCP Bidirectional Throughput - :header: "Command Used","am62dxx_evm-fs: THROUGHPUT (Mbits/sec)","am62dxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","304.93 (min 301.56, max 307.77)","11.34 (min 11.00, max 11.84)" - -.. rubric:: TCP Bidirectional Throughput Interrupt Pacing - :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing - -.. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing - :header: "Command Used","am62dxx_evm-fs: THROUGHPUT (Mbits/sec)","am62dxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","305.08 (min 301.65, max 308.15)","11.48 (min 11.33, max 11.63)" - -.. rubric:: UDP Throughput - :name: CPSW2g-udp-throughput-0-loss +UDP Throughput +^^^^^^^^^^^^^^ .. csv-table:: CPSW2g UDP Egress Throughput 0 loss :header: "Frame Size(bytes)","am62dxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62dxx_evm-fs: THROUGHPUT (Mbits/sec)","am62dxx_evm-fs: Packets Per Second (kPPS)","am62dxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","57.87 (min 55.77, max 59.88)","113.00 (min 109.00, max 117.00)","25.67 (min 25.08, max 26.03)" - "128","82.00","113.83 (min 109.02, max 116.77)","111.13 (min 106.00, max 114.00)","25.77 (min 25.07, max 26.02)" - "256","210.00","148.42 (min 63.90, max 160.50)","72.13 (min 31.00, max 78.00)","19.34 (min 6.07, max 21.69)" - "1024","978.00","181.49 (min 181.48, max 181.50)","22.00","7.98 (min 7.88, max 8.14)" - "1518","1472.00","179.37 (min 179.36, max 179.37)","15.00","7.84 (min 7.65, max 7.99)" + "64","","46.03 (min 45.82, max 46.24)","89.50 (min 89.00, max 90.00)","38.12 (min 38.01, max 38.23)" + "128","","91.09 (min 90.84, max 91.34)","89.00","38.04 (min 37.99, max 38.09)" .. csv-table:: CPSW2g UDP Ingress Throughput 0 loss :header: "Frame Size(bytes)","am62dxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62dxx_evm-fs: THROUGHPUT (Mbits/sec)","am62dxx_evm-fs: Packets Per Second (kPPS)","am62dxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","1.86 (min 1.48, max 2.36)","3.63 (min 3.00, max 5.00)","1.53 (min 0.53, max 3.81)" - "128","82.00","4.47 (min 4.40, max 4.71)","4.13 (min 4.00, max 5.00)","1.02 (min 0.76, max 1.65)" - "256","210.00","10.41 (min 10.03, max 10.85)","5.00","2.42 (min 0.94, max 5.27)" - "1024","978.00","43.19 (min 42.60, max 44.24)","5.00","2.08 (min 1.31, max 4.18)" - "1518","1472.00","62.15 (min 61.23, max 64.77)","5.11 (min 5.00, max 6.00)","2.90 (min 1.80, max 4.28)" + "64","","1.97 (min 1.54, max 2.15)","3.75 (min 3.00, max 4.00)","2.24 (min 1.52, max 3.73)" + "128","","5.15 (min 5.12, max 5.22)","5.00","2.97 (min 2.13, max 4.33)" + "256","","9.78 (min 9.01, max 10.65)","4.75 (min 4.00, max 5.00)","2.84 (min 1.03, max 5.42)" + "1024","","43.09 (min 38.50, max 51.61)","5.20 (min 5.00, max 6.00)","3.78 (min 2.73, max 6.02)" + "1518","","60.76 (min 55.35, max 64.77)","5.20 (min 5.00, max 6.00)","3.50 (min 2.30, max 5.87)" .. csv-table:: CPSW2g UDP Ingress Throughput possible loss :header: "Frame Size(bytes)","am62dxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62dxx_evm-fs: THROUGHPUT (Mbits/sec)","am62dxx_evm-fs: Packets Per Second (kPPS)","am62dxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)","am62dxx_evm-fs: Packet Loss %" - "64","18.00","82.68 (min 69.85, max 95.35)","161.50 (min 136.00, max 186.00)","40.23 (min 38.26, max 42.06)","31.04 (min 0.18, max 62.30)" - "128","82.00","174.59 (min 130.93, max 188.67)","170.50 (min 128.00, max 184.00)","41.26 (min 39.46, max 42.34)","44.65 (min 0.39, max 61.31)" - "256","210.00","320.34 (min 259.61, max 369.04)","156.50 (min 127.00, max 180.00)","40.94 (min 38.60, max 42.78)","24.24 (min 0.37, max 49.24)" - "1024","978.00","876.98 (min 837.06, max 913.30)","106.86 (min 102.00, max 111.00)","40.28 (min 39.04, max 41.61)","0.61 (min 0.18, max 1.17)" - "1518","1472.00","892.63 (min 786.50, max 934.06)","75.78 (min 67.00, max 79.00)","38.92 (min 34.39, max 40.76)","1.04 (min 0.25, max 2.13)" + "64","","75.95 (min 67.93, max 98.26)","148.50 (min 133.00, max 192.00)","36.17 (min 34.38, max 38.15)","17.56 (min 1.13, max 64.36)" + "128","","134.40 (min 129.62, max 137.51)","131.33 (min 127.00, max 134.00)","36.01 (min 33.95, max 38.93)","2.45 (min 0.78, max 4.73)" + "256","","304.24 (min 272.41, max 387.16)","148.50 (min 133.00, max 189.00)","37.26 (min 35.26, max 39.33)","29.49 (min 0.84, max 62.35)" + "1024","","909.90 (min 871.42, max 937.29)","111.00 (min 106.00, max 114.00)","38.69 (min 35.09, max 41.03)","0.77 (min 0.22, max 1.35)" + "1518","","909.09 (min 860.26, max 930.92)","77.20 (min 73.00, max 79.00)","38.12 (min 35.95, max 39.53)","0.39 (min 0.09, max 1.30)" | -Linux OSPI Flash Driver ------------------------ - -RAW - -.. csv-table:: OSPI Raw Flash Driver - :header: "File size (Mbytes)","am62dxx_evm-fs: Raw Read Throughput (Mbytes/sec)" - - "10","227.27" - "50","217.39" - -| - -USB Driver ----------- +USB +=== USB Device Controller -^^^^^^^^^^^^^^^^^^^^^ +--------------------- .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT :header: "Number of Blocks","am62dxx_evm-fs: Throughput (MB/sec)" - "150","31.79 (min 20.00, max 34.80)" + "150","31.76 (min 26.90, max 40.40)" .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT :header: "Number of Blocks","am62dxx_evm-fs: Throughput (MB/sec)" - "150","30.64 (min 24.70, max 32.80)" + "150","27.41 (min 22.40, max 33.30)" | -CRYPTO Driver -------------- +CRYPTO +====== OpenSSL Performance -^^^^^^^^^^^^^^^^^^^ +------------------- .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am62dxx_evm-fs: throughput (KBytes/Sec)" - "aes-128-cbc","1024","23511.81 (min 22199.30, max 24003.58)" - "aes-128-cbc","16","424.58 (min 400.86, max 435.84)" - "aes-128-cbc","16384","85633.71 (min 84454.06, max 86250.84)" - "aes-128-cbc","256","7031.72 (min 6708.74, max 7234.22)" - "aes-128-cbc","64","1847.05 (min 1762.45, max 1904.09)" - "aes-128-cbc","8192","72108.37 (min 70328.32, max 72742.23)" - "aes-128-ecb","1024","24118.53 (min 23312.38, max 24716.29)" - "aes-128-ecb","16","436.57 (min 412.31, max 446.68)" - "aes-128-ecb","16384","88354.13 (min 87086.42, max 89161.73)" - "aes-128-ecb","256","7186.42 (min 6870.70, max 7333.55)" - "aes-128-ecb","64","1899.44 (min 1787.95, max 1948.69)" - "aes-128-ecb","8192","74231.47 (min 72373.59, max 74986.84)" - "aes-192-cbc","1024","23054.25 (min 21801.64, max 23675.90)" - "aes-192-cbc","16","426.41 (min 403.57, max 437.48)" - "aes-192-cbc","16384","77032.11 (min 75956.22, max 77736.62)" - "aes-192-cbc","256","6940.00 (min 6559.74, max 7117.23)" - "aes-192-cbc","64","1855.78 (min 1761.34, max 1896.92)" - "aes-192-cbc","8192","66243.58 (min 64476.50, max 67100.67)" - "aes-192-ecb","1024","23512.23 (min 22242.65, max 24032.94)" - "aes-192-ecb","16","435.23 (min 416.22, max 443.56)" - "aes-192-ecb","16384","79489.02 (min 77490.86, max 80303.45)" - "aes-192-ecb","256","7143.25 (min 6823.34, max 7303.17)" - "aes-192-ecb","64","1892.40 (min 1786.77, max 1936.68)" - "aes-192-ecb","8192","67959.13 (min 66158.59, max 68741.80)" - "aes-256-cbc","1024","22439.42 (min 21377.71, max 22966.27)" - "aes-256-cbc","16","427.52 (min 403.20, max 438.70)" - "aes-256-cbc","16384","70787.07 (min 69331.63, max 71254.02)" - "aes-256-cbc","256","6881.53 (min 6456.32, max 7109.89)" - "aes-256-cbc","64","1843.97 (min 1729.77, max 1889.94)" - "aes-256-cbc","8192","61376.17 (min 59588.61, max 62046.21)" - "aes-256-ecb","1024","23099.73 (min 22211.93, max 23634.60)" - "aes-256-ecb","16","435.45 (min 415.85, max 443.94)" - "aes-256-ecb","16384","72642.56 (min 71390.55, max 73176.41)" - "aes-256-ecb","256","7103.46 (min 6742.61, max 7271.85)" - "aes-256-ecb","64","1893.37 (min 1771.86, max 1932.05)" - "aes-256-ecb","8192","62965.42 (min 61532.84, max 63539.88)" - "sha256","1024","38004.99 (min 37393.07, max 38651.90)" - "sha256","16","632.01 (min 623.18, max 639.92)" - "sha256","16384","300601.34 (min 296621.40, max 303781.21)" - "sha256","256","9924.28 (min 9777.24, max 10089.73)" - "sha256","64","2500.19 (min 2459.39, max 2530.41)" - "sha256","8192","203442.52 (min 201657.00, max 204671.66)" - "sha512","1024","26159.23 (min 25836.20, max 26417.49)" - "sha512","16","614.39 (min 601.95, max 625.40)" - "sha512","16384","68399.79 (min 68119.21, max 68577.96)" - "sha512","256","8720.94 (min 8533.85, max 8851.97)" - "sha512","64","2456.83 (min 2404.99, max 2503.42)" - "sha512","8192","61606.57 (min 61311.66, max 61794.99)" + "aes-128-cbc","1024","24568.83" + "aes-128-cbc","16","446.98" + "aes-128-cbc","16384","87168.34" + "aes-128-cbc","256","7377.41" + "aes-128-cbc","64","1969.19" + "aes-128-cbc","8192","73949.18" + "aes-128-ecb","1024","24998.57" + "aes-128-ecb","16","453.88" + "aes-128-ecb","16384","89085.27" + "aes-128-ecb","256","7518.29" + "aes-128-ecb","64","2004.20" + "aes-128-ecb","8192","75653.12" + "aes-192-cbc","1024","23895.38" + "aes-192-cbc","16","446.65" + "aes-192-cbc","16384","78091.61" + "aes-192-cbc","256","7297.54" + "aes-192-cbc","64","1949.01" + "aes-192-cbc","8192","67155.29" + "aes-192-ecb","1024","24523.09" + "aes-192-ecb","16","457.21" + "aes-192-ecb","16384","80439.98" + "aes-192-ecb","256","7495.51" + "aes-192-ecb","64","2025.96" + "aes-192-ecb","8192","69457.24" + "aes-256-cbc","1024","23421.95" + "aes-256-cbc","16","447.55" + "aes-256-cbc","16384","71516.16" + "aes-256-cbc","256","7258.20" + "aes-256-cbc","64","1949.67" + "aes-256-cbc","8192","61909.67" + "aes-256-ecb","1024","24072.53" + "aes-256-ecb","16","458.01" + "aes-256-ecb","16384","73820.84" + "aes-256-ecb","256","7437.40" + "aes-256-ecb","64","2014.89" + "aes-256-ecb","8192","64184.32" + "sha256","1024","32333.82" + "sha256","16","535.81" + "sha256","16384","277938.18" + "sha256","256","8409.69" + "sha256","64","2122.37" + "sha256","8192","183263.23" + "sha512","1024","26990.25" + "sha512","16","520.24" + "sha512","16384","110586.54" + "sha512","256","7864.32" + "sha512","64","2082.37" + "sha512","8192","91439.10" .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am62dxx_evm-fs: CPU Load" - "aes-128-cbc","31.50 (min 30.00, max 32.00)" - "aes-128-ecb","32.75 (min 31.00, max 34.00)" - "aes-192-cbc","31.50 (min 30.00, max 32.00)" - "aes-192-ecb","32.13 (min 30.00, max 33.00)" - "aes-256-cbc","31.00 (min 29.00, max 32.00)" - "aes-256-ecb","31.63 (min 30.00, max 33.00)" - "sha256","95.50 (min 94.00, max 96.00)" - "sha512","95.63 (min 95.00, max 96.00)" + "aes-128-cbc","32.00" + "aes-128-ecb","34.00" + "aes-192-cbc","32.00" + "aes-192-ecb","33.00" + "aes-256-cbc","32.00" + "aes-256-ecb","32.00" + "sha256","95.00" + "sha512","96.00" Listed for each algorithm are the code snippets used to run each benchmark test. @@ -582,3 +560,5 @@ Listed for each algorithm are the code snippets used to run each :: time -v openssl speed -elapsed -evp aes-128-cbc + +| diff --git a/source/devices/AM62LX/linux/Linux_Performance_Guide.rst b/source/devices/AM62LX/linux/Linux_Performance_Guide.rst index 0457fac3a..ced80fa4d 100644 --- a/source/devices/AM62LX/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62LX/linux/Linux_Performance_Guide.rst @@ -1,10 +1,10 @@ +################################# + Linux 12.00.00 Performance Guide +################################# -==================================== - Linux 11.02.08.02 Performance Guide -==================================== - -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -17,8 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -26,18 +27,20 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ +| +***************** System Benchmarks ------------------ +***************** + +| LMBench -^^^^^^^ +======= + LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at @@ -64,124 +67,125 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "af_unix_sock_stream_latency (microsec)","33.14 (min 32.70, max 34.50)" - "af_unix_socket_stream_bandwidth (mb\s)","458.31 (min 416.22, max 548.49)" - "bw_file_rd-io-1mb (mb/s)","971.18 (min 906.62, max 1128.85)" - "bw_file_rd-o2c-1mb (mb/s)","515.41 (min 508.13, max 525.95)" - "bw_mem-bcopy-16mb (mb/s)","742.11 (min 704.91, max 849.57)" - "bw_mem-bcopy-1mb (mb/s)","755.61 (min 687.76, max 942.68)" - "bw_mem-bcopy-2mb (mb/s)","748.93 (min 681.43, max 958.47)" - "bw_mem-bcopy-4mb (mb/s)","781.73 (min 727.40, max 944.73)" - "bw_mem-bcopy-8mb (mb/s)","801.94 (min 738.69, max 994.53)" - "bw_mem-bzero-16mb (mb/s)","2250.16 (min 2104.16, max 2805.54)" - "bw_mem-bzero-1mb (mb/s)","1502.54 (min 687.76, max 2800.63)" - "bw_mem-bzero-2mb (mb/s)","1499.05 (min 681.43, max 2807.02)" - "bw_mem-bzero-4mb (mb/s)","1516.48 (min 727.40, max 2806.03)" - "bw_mem-bzero-8mb (mb/s)","1526.92 (min 738.69, max 2807.02)" - "bw_mem-cp-16mb (mb/s)","422.87 (min 397.28, max 517.95)" - "bw_mem-cp-1mb (mb/s)","1450.86 (min 401.50, max 3062.20)" - "bw_mem-cp-2mb (mb/s)","1387.71 (min 399.24, max 2898.03)" - "bw_mem-cp-4mb (mb/s)","1383.65 (min 434.12, max 2868.93)" - "bw_mem-cp-8mb (mb/s)","1376.95 (min 436.56, max 2827.85)" - "bw_mem-fcp-16mb (mb/s)","751.75 (min 710.13, max 879.75)" - "bw_mem-fcp-1mb (mb/s)","1538.99 (min 777.24, max 2800.63)" - "bw_mem-fcp-2mb (mb/s)","1534.98 (min 762.92, max 2807.02)" - "bw_mem-fcp-4mb (mb/s)","1547.58 (min 788.72, max 2806.03)" - "bw_mem-fcp-8mb (mb/s)","1556.02 (min 804.67, max 2807.02)" - "bw_mem-frd-16mb (mb/s)","1328.40 (min 1227.75, max 1609.50)" - "bw_mem-frd-1mb (mb/s)","1057.89 (min 777.24, max 1525.94)" - "bw_mem-frd-2mb (mb/s)","1063.22 (min 762.92, max 1550.87)" - "bw_mem-frd-4mb (mb/s)","1080.32 (min 788.72, max 1555.81)" - "bw_mem-frd-8mb (mb/s)","1093.89 (min 804.67, max 1611.60)" - "bw_mem-fwr-16mb (mb/s)","2270.76 (min 2123.71, max 2819.38)" - "bw_mem-fwr-1mb (mb/s)","1876.87 (min 1218.45, max 3062.20)" - "bw_mem-fwr-2mb (mb/s)","1828.62 (min 1203.61, max 2898.03)" - "bw_mem-fwr-4mb (mb/s)","1809.91 (min 1247.47, max 2868.93)" - "bw_mem-fwr-8mb (mb/s)","1802.83 (min 1250.00, max 2827.85)" - "bw_mem-rd-16mb (mb/s)","1345.30 (min 1264.92, max 1616.65)" - "bw_mem-rd-1mb (mb/s)","953.97 (min 525.76, max 1599.43)" - "bw_mem-rd-2mb (mb/s)","940.68 (min 511.31, max 1592.61)" - "bw_mem-rd-4mb (mb/s)","978.01 (min 513.68, max 1601.71)" - "bw_mem-rd-8mb (mb/s)","1032.26 (min 681.02, max 1605.14)" - "bw_mem-rdwr-16mb (mb/s)","732.31 (min 663.63, max 858.23)" - "bw_mem-rdwr-1mb (mb/s)","503.57 (min 401.50, max 680.27)" - "bw_mem-rdwr-2mb (mb/s)","494.06 (min 399.24, max 692.28)" - "bw_mem-rdwr-4mb (mb/s)","553.67 (min 434.12, max 778.21)" - "bw_mem-rdwr-8mb (mb/s)","585.97 (min 436.56, max 857.72)" - "bw_mem-wr-16mb (mb/s)","763.43 (min 702.83, max 927.75)" - "bw_mem-wr-1mb (mb/s)","573.83 (min 525.76, max 686.34)" - "bw_mem-wr-2mb (mb/s)","559.79 (min 507.29, max 692.28)" - "bw_mem-wr-4mb (mb/s)","633.10 (min 513.68, max 800.24)" - "bw_mem-wr-8mb (mb/s)","708.78 (min 627.60, max 871.93)" - "bw_mmap_rd-mo-1mb (mb/s)","1245.57 (min 1156.29, max 1501.22)" - "bw_mmap_rd-o2c-1mb (mb/s)","509.91 (min 505.22, max 519.21)" - "bw_pipe (mb/s)","426.35 (min 382.91, max 518.60)" - "bw_unix (mb/s)","458.31 (min 416.22, max 548.49)" - "lat_connect (us)","72.81 (min 71.39, max 76.51)" - "lat_ctx-2-128k (us)","17.46 (min 11.52, max 25.12)" - "lat_ctx-2-256k (us)","31.42 (min 26.43, max 34.63)" - "lat_ctx-4-128k (us)","15.70 (min 12.41, max 17.36)" - "lat_ctx-4-256k (us)","8.13 (min 0.00, max 16.57)" - "lat_fs-0k (num_files)","198.17 (min 189.00, max 213.00)" - "lat_fs-10k (num_files)","84.33 (min 76.00, max 90.00)" - "lat_fs-1k (num_files)","127.00 (min 113.00, max 140.00)" - "lat_fs-4k (num_files)","128.17 (min 119.00, max 134.00)" - "lat_mem_rd-stride128-sz1000k (ns)","49.98 (min 41.71, max 52.46)" - "lat_mem_rd-stride128-sz125k (ns)","6.19 (min 6.16, max 6.22)" - "lat_mem_rd-stride128-sz250k (ns)","13.46 (min 6.99, max 21.17)" - "lat_mem_rd-stride128-sz31k (ns)","3.59 (min 2.42, max 4.45)" - "lat_mem_rd-stride128-sz50 (ns)","2.40 (min 2.40, max 2.41)" - "lat_mem_rd-stride128-sz500k (ns)","45.07 (min 38.07, max 48.16)" - "lat_mem_rd-stride128-sz62k (ns)","5.74 (min 5.72, max 5.75)" - "lat_mmap-1m (us)","59.83 (min 57.00, max 70.00)" + "af_unix_sock_stream_latency (microsec)","28.47 (min 25.57, max 29.86)" + "af_unix_socket_stream_bandwidth (mb\s)","473.68 (min 441.13, max 551.95)" + "bw_file_rd-io-1mb (mb/s)","952.82 (min 909.92, max 1066.29)" + "bw_file_rd-o2c-1mb (mb/s)","391.04 (min 55.18, max 547.95)" + "bw_mem-bcopy-16mb (mb/s)","749.66 (min 705.87, max 858.60)" + "bw_mem-bcopy-1mb (mb/s)","769.84 (min 694.44, max 909.09)" + "bw_mem-bcopy-2mb (mb/s)","752.70 (min 690.49, max 913.94)" + "bw_mem-bcopy-4mb (mb/s)","775.04 (min 704.72, max 954.65)" + "bw_mem-bcopy-8mb (mb/s)","809.57 (min 727.14, max 971.23)" + "bw_mem-bzero-16mb (mb/s)","2279.05 (min 2106.37, max 2725.72)" + "bw_mem-bzero-1mb (mb/s)","1526.26 (min 694.44, max 2725.29)" + "bw_mem-bzero-2mb (mb/s)","1515.47 (min 690.49, max 2727.45)" + "bw_mem-bzero-4mb (mb/s)","1526.69 (min 704.72, max 2724.33)" + "bw_mem-bzero-8mb (mb/s)","1546.56 (min 727.14, max 2725.72)" + "bw_mem-cp-16mb (mb/s)","426.86 (min 395.35, max 500.74)" + "bw_mem-cp-1mb (mb/s)","1439.36 (min 395.41, max 2878.53)" + "bw_mem-cp-2mb (mb/s)","1400.78 (min 386.25, max 2819.38)" + "bw_mem-cp-4mb (mb/s)","1398.29 (min 434.36, max 2764.34)" + "bw_mem-cp-8mb (mb/s)","1393.47 (min 432.85, max 2750.09)" + "bw_mem-fcp-16mb (mb/s)","751.74 (min 710.38, max 853.20)" + "bw_mem-fcp-1mb (mb/s)","1557.43 (min 776.04, max 2725.29)" + "bw_mem-fcp-2mb (mb/s)","1551.17 (min 764.33, max 2727.45)" + "bw_mem-fcp-4mb (mb/s)","1563.30 (min 790.28, max 2724.33)" + "bw_mem-fcp-8mb (mb/s)","1574.04 (min 798.08, max 2725.72)" + "bw_mem-frd-16mb (mb/s)","1329.96 (min 1253.62, max 1537.28)" + "bw_mem-frd-1mb (mb/s)","1049.23 (min 776.04, max 1494.21)" + "bw_mem-frd-2mb (mb/s)","1043.79 (min 764.33, max 1423.99)" + "bw_mem-frd-4mb (mb/s)","1062.05 (min 790.28, max 1526.14)" + "bw_mem-frd-8mb (mb/s)","1092.66 (min 798.08, max 1551.89)" + "bw_mem-fwr-16mb (mb/s)","2295.38 (min 2122.02, max 2738.79)" + "bw_mem-fwr-1mb (mb/s)","1857.24 (min 1178.25, max 2878.53)" + "bw_mem-fwr-2mb (mb/s)","1819.10 (min 1196.53, max 2819.38)" + "bw_mem-fwr-4mb (mb/s)","1803.17 (min 1166.35, max 2764.34)" + "bw_mem-fwr-8mb (mb/s)","1816.38 (min 1237.43, max 2750.09)" + "bw_mem-rd-16mb (mb/s)","1348.11 (min 1269.04, max 1574.80)" + "bw_mem-rd-1mb (mb/s)","973.75 (min 532.67, max 1533.45)" + "bw_mem-rd-2mb (mb/s)","954.80 (min 508.78, max 1548.95)" + "bw_mem-rd-4mb (mb/s)","965.06 (min 582.16, max 1563.11)" + "bw_mem-rd-8mb (mb/s)","1028.38 (min 671.03, max 1574.49)" + "bw_mem-rdwr-16mb (mb/s)","742.83 (min 671.11, max 871.98)" + "bw_mem-rdwr-1mb (mb/s)","507.25 (min 395.41, max 758.58)" + "bw_mem-rdwr-2mb (mb/s)","508.60 (min 386.25, max 742.12)" + "bw_mem-rdwr-4mb (mb/s)","554.93 (min 434.36, max 758.58)" + "bw_mem-rdwr-8mb (mb/s)","594.72 (min 432.85, max 831.51)" + "bw_mem-wr-16mb (mb/s)","761.79 (min 696.29, max 875.47)" + "bw_mem-wr-1mb (mb/s)","597.49 (min 503.36, max 758.58)" + "bw_mem-wr-2mb (mb/s)","580.37 (min 508.78, max 742.12)" + "bw_mem-wr-4mb (mb/s)","649.66 (min 582.16, max 809.88)" + "bw_mem-wr-8mb (mb/s)","712.22 (min 660.23, max 831.51)" + "bw_mmap_rd-mo-1mb (mb/s)","1186.31 (min 1095.29, max 1369.06)" + "bw_mmap_rd-o2c-1mb (mb/s)","504.63 (min 466.13, max 534.00)" + "bw_pipe (mb/s)","397.94 (min 332.79, max 502.07)" + "bw_unix (mb/s)","473.68 (min 441.13, max 551.95)" + "lat_connect (us)","110.85 (min 108.39, max 113.45)" + "lat_ctx-2-128k (us)","40.79 (min 10.02, max 53.07)" + "lat_ctx-2-256k (us)","122.00 (min 19.85, max 160.05)" + "lat_ctx-4-128k (us)","79.14 (min 11.48, max 104.99)" + "lat_ctx-4-256k (us)","95.97 (min 8.50, max 179.58)" + "lat_fs-0k (num_files)","203.25 (min 195.00, max 212.00)" + "lat_fs-10k (num_files)","89.00 (min 81.00, max 96.00)" + "lat_fs-1k (num_files)","134.25 (min 118.00, max 146.00)" + "lat_fs-4k (num_files)","128.25 (min 120.00, max 136.00)" + "lat_mem_rd-stride128-sz1000k (ns)","49.87 (min 42.84, max 52.35)" + "lat_mem_rd-stride128-sz125k (ns)","6.18 (min 6.17, max 6.18)" + "lat_mem_rd-stride128-sz250k (ns)","11.22 (min 7.65, max 14.89)" + "lat_mem_rd-stride128-sz31k (ns)","3.16 (min 2.41, max 3.90)" + "lat_mem_rd-stride128-sz50 (ns)","2.41" + "lat_mem_rd-stride128-sz500k (ns)","45.38 (min 39.37, max 47.67)" + "lat_mem_rd-stride128-sz62k (ns)","5.74 (min 5.72, max 5.77)" + "lat_mmap-1m (us)","65.00 (min 57.00, max 68.00)" "lat_ops-double-add (ns)","3.21" - "lat_ops-double-div (ns)","17.65 (min 17.64, max 17.66)" - "lat_ops-double-mul (ns)","3.21 (min 3.21, max 3.22)" + "lat_ops-double-div (ns)","17.66 (min 17.65, max 17.66)" + "lat_ops-double-mul (ns)","3.21" "lat_ops-float-add (ns)","3.21" - "lat_ops-float-div (ns)","10.43 (min 10.42, max 10.43)" - "lat_ops-float-mul (ns)","3.21" + "lat_ops-float-div (ns)","10.44 (min 10.43, max 10.45)" + "lat_ops-float-mul (ns)","3.21 (min 3.21, max 3.22)" "lat_ops-int-add (ns)","0.80" "lat_ops-int-bit (ns)","0.53" "lat_ops-int-div (ns)","4.82 (min 4.81, max 4.82)" "lat_ops-int-mod (ns)","5.08 (min 5.08, max 5.09)" - "lat_ops-int-mul (ns)","3.45 (min 3.44, max 3.46)" + "lat_ops-int-mul (ns)","3.49" "lat_ops-int64-add (ns)","0.80" - "lat_ops-int64-bit (ns)","0.53 (min 0.53, max 0.54)" - "lat_ops-int64-div (ns)","7.62" - "lat_ops-int64-mod (ns)","5.88 (min 5.88, max 5.89)" - "lat_ops-int64-mul (ns)","3.99 (min 3.98, max 4.04)" - "lat_pagefault (us)","0.93 (min 0.81, max 0.98)" - "lat_pipe (us)","26.36 (min 24.71, max 27.33)" - "lat_proc-exec (us)","1447.93 (min 1267.60, max 1563.25)" - "lat_proc-fork (us)","1269.68 (min 1094.20, max 1328.60)" + "lat_ops-int64-bit (ns)","0.54 (min 0.53, max 0.54)" + "lat_ops-int64-div (ns)","7.63 (min 7.62, max 7.63)" + "lat_ops-int64-mod (ns)","5.89 (min 5.88, max 5.92)" + "lat_ops-int64-mul (ns)","3.98 (min 3.97, max 3.99)" + "lat_pagefault (us)","0.87 (min 0.77, max 0.93)" + "lat_pipe (us)","22.18 (min 20.56, max 23.06)" + "lat_proc-exec (us)","1469.13 (min 1298.00, max 1543.25)" + "lat_proc-fork (us)","1296.60 (min 1160.40, max 1372.00)" "lat_proc-proccall (us)","0.01" - "lat_select (us)","38.21 (min 38.12, max 38.41)" - "lat_sem (us)","4.12 (min 3.74, max 4.31)" - "lat_sig-catch (us)","6.22 (min 6.07, max 6.38)" - "lat_sig-install (us)","0.74 (min 0.72, max 0.78)" - "lat_sig-prot (us)","0.69 (min 0.49, max 0.88)" - "lat_syscall-fstat (us)","2.20 (min 2.12, max 2.31)" - "lat_syscall-null (us)","0.51 (min 0.51, max 0.52)" - "lat_syscall-open (us)","403.50 (min 362.00, max 459.33)" - "lat_syscall-read (us)","0.91 (min 0.90, max 0.91)" - "lat_syscall-stat (us)","5.41 (min 5.23, max 5.60)" - "lat_syscall-write (us)","0.86 (min 0.84, max 0.88)" - "lat_tcp (us)","1.02" - "lat_unix (us)","33.14 (min 32.70, max 34.50)" - "latency_for_0.50_mb_block_size (nanosec)","45.07 (min 38.07, max 48.16)" - "latency_for_1.00_mb_block_size (nanosec)","24.99 (min 0.00, max 52.46)" - "pipe_bandwidth (mb\s)","426.35 (min 382.91, max 518.60)" - "pipe_latency (microsec)","26.36 (min 24.71, max 27.33)" + "lat_select (us)","35.38 (min 35.29, max 35.56)" + "lat_sem (us)","3.55 (min 3.33, max 3.84)" + "lat_sig-catch (us)","6.24 (min 6.16, max 6.34)" + "lat_sig-install (us)","0.66 (min 0.65, max 0.70)" + "lat_sig-prot (us)","0.95 (min 0.88, max 1.01)" + "lat_syscall-fstat (us)","2.02 (min 1.97, max 2.12)" + "lat_syscall-null (us)","0.42 (min 0.41, max 0.42)" + "lat_syscall-open (us)","485.80 (min 413.85, max 568.44)" + "lat_syscall-read (us)","0.83 (min 0.82, max 0.83)" + "lat_syscall-stat (us)","4.74 (min 4.69, max 4.79)" + "lat_syscall-write (us)","0.75" + "lat_tcp (us)","0.86" + "lat_unix (us)","28.47 (min 25.57, max 29.86)" + "latency_for_0.50_mb_block_size (nanosec)","45.38 (min 39.37, max 47.67)" + "latency_for_1.00_mb_block_size (nanosec)","22.16 (min 0.00, max 52.35)" + "pipe_bandwidth (mb\s)","397.94 (min 332.79, max 502.07)" + "pipe_latency (microsec)","22.18 (min 20.56, max 23.06)" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","38.21 (min 38.12, max 38.41)" - "semaphore_latency (microsec)","4.12 (min 3.74, max 4.31)" - "signal_handler_latency (microsec)","0.74 (min 0.72, max 0.78)" - "signal_handler_overhead (microsec)","6.22 (min 6.07, max 6.38)" - "tcp_ip_connection_cost_to_localhost (microsec)","72.81 (min 71.39, max 76.51)" - "tcp_latency_using_localhost (microsec)","1.02" + "select_on_200_tcp_fds (microsec)","35.38 (min 35.29, max 35.56)" + "semaphore_latency (microsec)","3.55 (min 3.33, max 3.84)" + "signal_handler_latency (microsec)","0.66 (min 0.65, max 0.70)" + "signal_handler_overhead (microsec)","6.24 (min 6.16, max 6.34)" + "tcp_ip_connection_cost_to_localhost (microsec)","110.85 (min 108.39, max 113.45)" + "tcp_latency_using_localhost (microsec)","0.86" Dhrystone -^^^^^^^^^ +========= + Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. @@ -198,12 +202,13 @@ Execute the benchmark with the following: .. csv-table:: Dhrystone Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "cpu_clock (mhz)","400.00" - "dhrystone_per_mhz (dmips/mhz)","8.90" + "cpu_clock (mhz)","1250.00" + "dhrystone_per_mhz (dmips/mhz)","2.80" "dhrystone_per_second (dhrystonep)","6250000.00" Whetstone -^^^^^^^^^ +========= + Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -215,20 +220,22 @@ Execute the benchmark with the following: .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "whetstone (mips)","5833.33 (min 5000.00, max 10000.00)" + "whetstone (mips)","5000.00" Linpack -^^^^^^^ +======= + Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "linpack (kflops)","515149.67 (min 514088.00, max 516563.00)" + "linpack (kflops)","516120.50 (min 514627.00, max 517614.00)" NBench -^^^^^^ +====== + NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at @@ -238,18 +245,19 @@ https://nbench.io/articles/index.html .. csv-table:: NBench Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "assignment (iterations)","12.95 (min 12.90, max 12.97)" - "fourier (iterations)","20382.00 (min 20379.00, max 20384.00)" - "fp_emulation (iterations)","192.40 (min 192.33, max 192.47)" - "huffman (iterations)","1057.02 (min 1056.80, max 1057.20)" - "idea (iterations)","3075.14 (min 3074.90, max 3075.50)" - "lu_decomposition (iterations)","473.43 (min 470.10, max 479.32)" - "neural_net (iterations)","7.73 (min 7.72, max 7.73)" - "numeric_sort (iterations)","558.36 (min 555.56, max 562.43)" - "string_sort (iterations)","146.38 (min 146.36, max 146.39)" + "assignment (iterations)","12.64 (min 12.62, max 12.67)" + "fourier (iterations)","18560.75 (min 18556.00, max 18568.00)" + "fp_emulation (iterations)","182.43 (min 182.35, max 182.50)" + "huffman (iterations)","1063.13 (min 1063.00, max 1063.20)" + "idea (iterations)","2738.40 (min 2738.10, max 2738.60)" + "lu_decomposition (iterations)","473.09 (min 467.17, max 475.75)" + "neural_net (iterations)","8.63 (min 8.59, max 8.64)" + "numeric_sort (iterations)","483.39 (min 482.54, max 484.16)" + "string_sort (iterations)","150.19 (min 150.17, max 150.21)" Stream -^^^^^^ +====== + STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -269,50 +277,51 @@ Execute the benchmark with the following: .. csv-table:: Stream Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "add (mb/s)","1651.55 (min 1501.10, max 1914.10)" - "copy (mb/s)","1735.32 (min 1565.10, max 2033.20)" - "scale (mb/s)","1830.72 (min 1641.00, max 2134.40)" - "triad (mb/s)","1632.43 (min 1491.50, max 1888.40)" + "add (mb/s)","1523.87 (min 1498.10, max 1538.30)" + "copy (mb/s)","1629.27 (min 1610.40, max 1639.00)" + "scale (mb/s)","1672.47 (min 1632.10, max 1713.80)" + "triad (mb/s)","1512.27 (min 1488.30, max 1526.00)" CoreMarkPro -^^^^^^^^^^^ +=========== + CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, adding comprehensive support for multicore technology, a combination of integer and floating-point workloads, and data sets for utilizing larger memory subsystems. - .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" "cjpeg-rose7-preset (workloads/)","37.22 (min 37.04, max 37.31)" "core (workloads/)","0.27" - "coremark-pro ()","705.20 (min 686.04, max 730.09)" - "linear_alg-mid-100x100-sp (workloads/)","13.09 (min 13.06, max 13.11)" - "loops-all-mid-10k-sp (workloads/)","0.57 (min 0.56, max 0.60)" - "nnet_test (workloads/)","0.97" - "parser-125k (workloads/)","6.35 (min 5.62, max 6.99)" - "radix2-big-64k (workloads/)","19.56 (min 17.06, max 22.07)" - "sha-test (workloads/)","71.11 (min 69.93, max 72.46)" - "zip-test (workloads/)","18.99 (min 18.52, max 19.61)" + "coremark-pro ()","719.19 (min 697.16, max 733.93)" + "linear_alg-mid-100x100-sp (workloads/)","13.08" + "loops-all-mid-10k-sp (workloads/)","0.59 (min 0.56, max 0.60)" + "nnet_test (workloads/)","0.96" + "parser-125k (workloads/)","7.62 (min 7.04, max 8.06)" + "radix2-big-64k (workloads/)","19.26 (min 16.30, max 21.85)" + "sha-test (workloads/)","71.11 (min 69.44, max 72.46)" + "zip-test (workloads/)","19.04 (min 17.86, max 20.41)" .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "cjpeg-rose7-preset (workloads/)","72.89 (min 71.94, max 74.07)" - "core (workloads/)","0.54 (min 0.53, max 0.54)" - "coremark-pro ()","1226.57 (min 1185.87, max 1277.14)" - "linear_alg-mid-100x100-sp (workloads/)","26.03 (min 25.97, max 26.11)" - "loops-all-mid-10k-sp (workloads/)","1.03 (min 1.00, max 1.07)" - "nnet_test (workloads/)","1.93 (min 1.93, max 1.94)" - "parser-125k (workloads/)","6.48 (min 5.81, max 7.33)" - "radix2-big-64k (workloads/)","28.39 (min 25.03, max 31.84)" - "sha-test (workloads/)","140.47 (min 138.89, max 142.86)" - "zip-test (workloads/)","34.17 (min 32.26, max 35.71)" + "cjpeg-rose7-preset (workloads/)","72.20 (min 71.94, max 72.46)" + "core (workloads/)","0.53 (min 0.53, max 0.54)" + "coremark-pro ()","1209.22 (min 1186.40, max 1232.04)" + "linear_alg-mid-100x100-sp (workloads/)","26.04 (min 26.03, max 26.06)" + "loops-all-mid-10k-sp (workloads/)","0.99 (min 0.98, max 1.00)" + "nnet_test (workloads/)","1.91 (min 1.91, max 1.92)" + "parser-125k (workloads/)","7.61 (min 5.97, max 9.26)" + "radix2-big-64k (workloads/)","24.15 (min 23.08, max 25.21)" + "sha-test (workloads/)","137.01 (min 135.14, max 138.89)" + "zip-test (workloads/)","33.37 (min 32.26, max 34.48)" MultiBench -^^^^^^^^^^ +========== + MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -331,52 +340,54 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am62lxx_evm-fs: perf" - "4m-check (workloads/)","288.19 (min 268.24, max 314.90)" - "4m-check-reassembly (workloads/)","57.34 (min 52.49, max 63.94)" - "4m-check-reassembly-tcp (workloads/)","36.65 (min 34.01, max 40.39)" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","16.34 (min 15.28, max 18.05)" - "4m-check-reassembly-tcp-x264w2 (workloads/)","0.93 (min 0.91, max 0.95)" - "4m-cmykw2 (workloads/)","109.44 (min 107.88, max 112.05)" - "4m-cmykw2-rotatew2 (workloads/)","21.20 (min 19.46, max 23.21)" - "4m-reassembly (workloads/)","52.91 (min 49.41, max 58.24)" - "4m-rotatew2 (workloads/)","21.26 (min 19.25, max 23.63)" - "4m-tcp-mixed (workloads/)","95.76 (min 93.02, max 98.16)" - "4m-x264w2 (workloads/)","0.94 (min 0.93, max 0.96)" + "4m-check (workloads/)","287.19 (min 265.48, max 320.97)" + "4m-check-reassembly (workloads/)","66.74 (min 51.05, max 83.26)" + "4m-check-reassembly-tcp (workloads/)","41.39 (min 33.69, max 47.80)" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","16.75 (min 15.29, max 18.62)" + "4m-check-reassembly-tcp-x264w2 (workloads/)","0.87 (min 0.78, max 0.95)" + "4m-cmykw2 (workloads/)","107.63 (min 106.10, max 110.01)" + "4m-cmykw2-rotatew2 (workloads/)","22.14 (min 18.79, max 26.33)" + "4m-reassembly (workloads/)","74.80 (min 47.60, max 98.62)" + "4m-rotatew2 (workloads/)","24.65 (min 19.73, max 28.48)" + "4m-tcp-mixed (workloads/)","102.76 (min 98.16, max 108.84)" + "4m-x264w2 (workloads/)","0.93 (min 0.92, max 0.96)" "empty-wld (workloads/)","1.00" - "idct-4m (workloads/)","16.97 (min 16.90, max 17.08)" - "idct-4mw1 (workloads/)","16.97 (min 16.89, max 17.09)" - "ippktcheck-4m (workloads/)","288.16 (min 267.84, max 314.11)" - "ippktcheck-4mw1 (workloads/)","287.54 (min 267.58, max 312.77)" - "ipres-4m (workloads/)","68.41 (min 62.50, max 76.69)" - "ipres-4mw1 (workloads/)","68.31 (min 62.47, max 76.18)" - "md5-4m (workloads/)","18.99 (min 17.91, max 20.68)" - "md5-4mw1 (workloads/)","19.11 (min 18.16, max 20.15)" - "rgbcmyk-4m (workloads/)","56.70 (min 56.45, max 57.08)" - "rgbcmyk-4mw1 (workloads/)","56.64 (min 56.32, max 57.01)" - "rotate-4ms1 (workloads/)","19.37 (min 17.85, max 20.73)" - "rotate-4ms1w1 (workloads/)","19.04 (min 17.85, max 20.74)" - "rotate-4ms64 (workloads/)","19.51 (min 18.04, max 22.62)" - "rotate-4ms64w1 (workloads/)","19.33 (min 18.01, max 21.44)" + "idct-4m (workloads/)","17.23 (min 16.47, max 17.62)" + "idct-4mw1 (workloads/)","17.23 (min 16.48, max 17.62)" + "ippktcheck-4m (workloads/)","288.80 (min 267.09, max 323.63)" + "ippktcheck-4mw1 (workloads/)","288.26 (min 265.73, max 323.71)" + "ipres-4m (workloads/)","67.60 (min 60.10, max 78.25)" + "ipres-4mw1 (workloads/)","67.64 (min 59.78, max 78.21)" + "md5-4m (workloads/)","19.49 (min 18.37, max 20.95)" + "md5-4mw1 (workloads/)","19.53 (min 18.49, max 21.01)" + "rgbcmyk-4m (workloads/)","57.07 (min 56.69, max 57.47)" + "rgbcmyk-4mw1 (workloads/)","57.04 (min 56.64, max 57.29)" + "rotate-4ms1 (workloads/)","20.24 (min 17.49, max 23.21)" + "rotate-4ms1w1 (workloads/)","20.27 (min 17.48, max 23.20)" + "rotate-4ms64 (workloads/)","20.37 (min 17.64, max 23.22)" + "rotate-4ms64w1 (workloads/)","20.36 (min 17.62, max 23.29)" "x264-4mq (workloads/)","0.50 (min 0.50, max 0.51)" "x264-4mqw1 (workloads/)","0.50 (min 0.49, max 0.51)" +| + Boot-time Measurement ---------------------- +===================== Boot media: MMCSD -^^^^^^^^^^^^^^^^^ +----------------- .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am62lxx_evm-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","19.01 (min 17.68, max 27.66)" + "Linux boot time from SD with default rootfs (20 boot cycles)","21.87 (min 19.45, max 24.14)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. | ALSA SoC Audio Driver ---------------------- +===================== #. Access type - RW\_INTERLEAVED #. Channels - 2 @@ -386,34 +397,35 @@ ALSA SoC Audio Driver .. csv-table:: Audio Capture :header: "Sampling Rate (Hz)","am62lxx_evm-fs: Throughput (bits/sec)","am62lxx_evm-fs: CPU Load (%)" - "11025","352795.83 (min 352795.00, max 352797.00)","0.38 (min 0.29, max 0.48)" - "16000","511995.00 (min 511994.00, max 511996.00)","0.35 (min 0.20, max 0.79)" - "22050","705589.67 (min 705586.00, max 705594.00)","0.40 (min 0.30, max 0.53)" - "24000","705591.50 (min 705589.00, max 705595.00)","0.47 (min 0.37, max 0.60)" - "32000","1023987.83 (min 1023985.00, max 1023991.00)","0.30 (min 0.21, max 0.40)" - "44100","1411182.17 (min 1411179.00, max 1411186.00)","0.59 (min 0.50, max 0.72)" - "48000","1535980.33 (min 1535976.00, max 1535985.00)","0.91 (min 0.23, max 1.56)" - "88200","2822355.00 (min 2822349.00, max 2822363.00)","1.09 (min 0.93, max 1.20)" - "96000","3071938.67 (min 3071934.00, max 3071943.00)","1.22 (min 0.39, max 3.19)" + "11025","352795.00 (min 352794.00, max 352796.00)","0.36 (min 0.29, max 0.42)" + "16000","511994.00","0.37 (min 0.31, max 0.42)" + "22050","705588.00 (min 705587.00, max 705589.00)","0.36 (min 0.35, max 0.36)" + "24000","705590.00 (min 705589.00, max 705591.00)","0.38 (min 0.36, max 0.40)" + "32000","1023985.50 (min 1023984.00, max 1023987.00)","0.22 (min 0.21, max 0.23)" + "44100","1411180.00 (min 1411179.00, max 1411181.00)","0.54 (min 0.53, max 0.54)" + "48000","1535978.50 (min 1535977.00, max 1535980.00)","0.59 (min 0.19, max 0.99)" + "88200","2822351.50 (min 2822349.00, max 2822354.00)","0.83 (min 0.81, max 0.84)" + "96000","3071928.50 (min 3071926.00, max 3071931.00)","1.18 (min 0.39, max 1.96)" .. csv-table:: Audio Playback :header: "Sampling Rate (Hz)","am62lxx_evm-fs: Throughput (bits/sec)","am62lxx_evm-fs: CPU Load (%)" - "11025","352943.17 (min 352943.00, max 352944.00)","0.30 (min 0.19, max 0.42)" - "16000","512208.83 (min 512208.00, max 512209.00)","0.26 (min 0.17, max 0.42)" - "22050","705883.83 (min 705882.00, max 705888.00)","0.30 (min 0.25, max 0.41)" - "24000","705886.17 (min 705886.00, max 705887.00)","0.33 (min 0.29, max 0.36)" - "32000","1024415.50 (min 1024414.00, max 1024417.00)","0.30 (min 0.19, max 0.65)" - "44100","1411772.17 (min 1411772.00, max 1411773.00)","0.41 (min 0.38, max 0.45)" - "48000","1536622.00 (min 1536621.00, max 1536623.00)","0.50 (min 0.19, max 0.95)" - "88200","2823534.17 (min 2823531.00, max 2823536.00)","0.77 (min 0.73, max 0.81)" - "96000","3073226.00 (min 3073218.00, max 3073234.00)","0.56 (min 0.31, max 1.58)" + "11025","352943.67 (min 352943.00, max 352944.00)","0.45 (min 0.31, max 0.63)" + "16000","512209.00 (min 512208.00, max 512210.00)","0.33 (min 0.22, max 0.46)" + "22050","705886.67 (min 705885.00, max 705890.00)","0.36 (min 0.30, max 0.48)" + "24000","705888.00 (min 705887.00, max 705890.00)","0.42 (min 0.36, max 0.53)" + "32000","1024417.67 (min 1024416.00, max 1024420.00)","0.54 (min 0.24, max 1.00)" + "44100","1411774.67 (min 1411772.00, max 1411778.00)","0.52 (min 0.42, max 0.67)" + "48000","1536624.67 (min 1536622.00, max 1536628.00)","0.48 (min 0.33, max 0.77)" + "88200","2823538.00 (min 2823532.00, max 2823546.00)","0.74 (min 0.62, max 0.91)" + "96000","3073210.00 (min 3073190.00, max 3073224.00)","0.48 (min 0.37, max 0.55)" | Ethernet ------------------ -Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html +======== + +Ethernet performance benchmarks were measured using :command:`netperf` 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, @@ -423,11 +435,11 @@ CPU utilization was captured as the total percentage used across all cores on th while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. -In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth +In this scenario, :command:`netperf` options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: -:: +.. code-block:: console burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) burst_size = 500000000 / 8 / 1472 / 100 = 425 @@ -435,21 +447,21 @@ no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B data wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the :command:`netperf` test with no bandwidth limit (remove -b/-w options). -In order to start a netperf client on one device, the other device must have netserver running. -To start netserver: +In order to start a :command:`netperf` client on one device, the other device must have :command:`netserver` running. +To start :command:`netserver`: -:: +.. code-block:: console netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger :command:`netperf` clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test. -:: +.. code-block:: console #!/bin/bash for i in 1 @@ -461,100 +473,77 @@ additional timing measurements during the test. -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for +Running the following commands will trigger :command:`netperf` clients to measure UDP burst performance for 60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run :command:`netperf` client from DUT and start :command:`netserver` on tester. -:: +.. code-block:: console netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run :command:`netperf` client from tester and start :command:`netserver` on DUT. -:: +.. code-block:: console netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE +CPSW/CPSW2g/CPSW3g Ethernet +--------------------------- -CPSW/CPSW2g/CPSW3g Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -- CPSW2g: AM65x, J7200, J721e, J721S2, J784S4, J742S2 -- CPSW3g: AM64x, AM62x, AM62ax, AM62px - -.. rubric:: TCP Bidirectional Throughput - :name: CPSW2g-tcp-bidirectional-throughput +TCP Bidirectional Throughput +^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am62lxx_evm-fs: THROUGHPUT (Mbits/sec)","am62lxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1250.65 (min 1162.63, max 1320.05)","99.45 (min 98.50, max 99.91)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1048.02 (min 1035.65, max 1060.38)","99.10 (min 98.47, max 99.72)" -.. rubric:: TCP Bidirectional Throughput Interrupt Pacing - :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing +TCP Bidirectional Throughput Interrupt Pacing +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am62lxx_evm-fs: THROUGHPUT (Mbits/sec)","am62lxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1276.53 (min 1096.32, max 1414.15)","98.53 (min 97.46, max 99.97)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1182.95 (min 1179.68, max 1186.21)","95.50 (min 95.39, max 95.60)" -.. rubric:: UDP Throughput - :name: CPSW2g-udp-throughput-0-loss +UDP Throughput +^^^^^^^^^^^^^^ .. csv-table:: CPSW2g UDP Egress Throughput 0 loss :header: "Frame Size(bytes)","am62lxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62lxx_evm-fs: THROUGHPUT (Mbits/sec)","am62lxx_evm-fs: Packets Per Second (kPPS)","am62lxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","48.02 (min 45.73, max 49.33)","93.57 (min 89.00, max 96.00)","81.68 (min 79.75, max 83.43)" - "128","82.00","91.45 (min 88.25, max 95.00)","89.29 (min 86.00, max 93.00)","76.38 (min 50.57, max 81.56)" - "256","210.00","173.95 (min 160.08, max 184.15)","85.00 (min 78.00, max 90.00)","76.66 (min 59.01, max 80.28)" - "1024","978.00","470.51 (min 73.73, max 704.77)","57.50 (min 9.00, max 86.00)","56.11 (min 7.01, max 79.70)" - "1518","1472.00","666.11 (min 646.32, max 702.83)","54.71 (min 53.00, max 58.00)","72.76 (min 71.76, max 74.10)" + "64","","36.89 (min 36.82, max 36.96)","72.00","75.48 (min 75.29, max 75.67)" + "128","","80.20","78.00","77.85" + "256","","145.24","71.00","75.58" + "1024","","355.15 (min 137.62, max 572.68)","43.50 (min 17.00, max 70.00)","49.89 (min 24.19, max 75.58)" + "1518","","623.50 (min 619.61, max 627.38)","51.50 (min 51.00, max 52.00)","72.61 (min 71.63, max 73.58)" .. csv-table:: CPSW2g UDP Ingress Throughput 0 loss :header: "Frame Size(bytes)","am62lxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62lxx_evm-fs: THROUGHPUT (Mbits/sec)","am62lxx_evm-fs: Packets Per Second (kPPS)","am62lxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","2.28 (min 2.25, max 2.36)","4.14 (min 4.00, max 5.00)","2.97 (min 1.94, max 6.96)" - "128","82.00","4.96 (min 4.40, max 5.43)","4.86 (min 4.00, max 5.00)","3.14 (min 2.03, max 4.94)" - "256","210.00","9.04 (min 1.02, max 10.85)","4.17 (min 0.00, max 5.00)","3.31 (min 0.98, max 6.33)" - "1024","978.00","38.40 (min 6.55, max 43.42)","4.50 (min 1.00, max 5.00)","4.55 (min 1.23, max 8.23)" - "1518","1472.00","52.60 (min 4.71, max 62.41)","4.17 (min 0.00, max 5.00)","5.17 (min 0.70, max 9.26)" + "64","","1.54 (min 1.48, max 1.59)","3.00","1.97 (min 1.84, max 2.10)" + "128","","4.46 (min 4.40, max 4.51)","4.00","3.70 (min 2.67, max 4.73)" + "256","","10.14 (min 10.04, max 10.24)","5.00","3.06 (min 3.00, max 3.11)" + "1024","","42.19 (min 41.78, max 42.60)","5.00","3.49 (min 3.40, max 3.57)" + "1518","","34.74 (min 8.24, max 61.23)","3.00 (min 1.00, max 5.00)","3.65 (min 1.02, max 6.27)" .. csv-table:: CPSW2g UDP Ingress Throughput possible loss :header: "Frame Size(bytes)","am62lxx_evm-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62lxx_evm-fs: THROUGHPUT (Mbits/sec)","am62lxx_evm-fs: Packets Per Second (kPPS)","am62lxx_evm-fs: CPU Load % (LOCAL_CPU_UTIL)","am62lxx_evm-fs: Packet Loss %" - "64","18.00","72.96 (min 69.39, max 76.56)","142.57 (min 136.00, max 150.00)","82.64 (min 81.65, max 85.19)","72.14 (min 61.99, max 82.53)" - "128","82.00","144.83 (min 137.49, max 154.44)","141.43 (min 134.00, max 151.00)","85.13 (min 83.80, max 86.33)","68.86 (min 55.17, max 78.22)" - "256","210.00","280.59 (min 262.68, max 303.33)","137.00 (min 128.00, max 148.00)","85.01 (min 82.70, max 87.15)","49.21 (min 31.44, max 66.24)" - "1024","978.00","809.96 (min 576.67, max 890.96)","99.00 (min 70.00, max 109.00)","88.52 (min 84.69, max 92.80)","7.02 (min 4.65, max 10.64)" - "1518","1472.00","787.28 (min 704.86, max 853.41)","66.83 (min 60.00, max 72.00)","82.25 (min 73.30, max 87.11)","6.83 (min 2.70, max 10.83)" + "64","","47.26 (min 46.91, max 47.61)","92.50 (min 92.00, max 93.00)","65.65 (min 64.95, max 66.34)","79.50 (min 79.03, max 79.97)" + "128","","89.46 (min 88.85, max 90.07)","87.50 (min 87.00, max 88.00)","67.22 (min 67.01, max 67.42)","83.49 (min 80.73, max 86.25)" + "256","","187.89 (min 178.98, max 196.80)","91.50 (min 87.00, max 96.00)","67.38 (min 65.87, max 68.89)","55.23 (min 51.03, max 59.42)" + "1024","","590.71 (min 544.95, max 636.47)","72.50 (min 67.00, max 78.00)","74.71 (min 70.00, max 79.41)","21.13 (min 16.49, max 25.77)" + "1518","","784.17 (min 775.84, max 792.50)","66.50 (min 66.00, max 67.00)","76.55 (min 75.73, max 77.37)","14.76 (min 12.32, max 17.20)" -Linux OSPI Flash Driver ------------------------ - -.. rubric:: UBIFS - :name: am62lxx-evm-ospi-ubifs - -.. csv-table:: OSPI Flash Driver - :header: "Buffer size (bytes)","am62lxx_evm-fs: Write UBIFS Throughput (Mbytes/sec)","am62lxx_evm-fs: Write UBIFS CPU Load (%)","am62lxx_evm-fs: Read UBIFS Throughput (Mbytes/sec)","am62lxx_evm-fs: Read UBIFS CPU Load (%)" - - "102400","3.84 (min 3.55, max 4.17)","29.68 (min 24.49, max 34.58)","11.32 (min 11.19, max 11.63)","33.47 (min 31.25, max 35.29)" - "262144","3.77 (min 3.49, max 4.27)","28.92 (min 24.49, max 33.65)","11.31 (min 11.20, max 11.55)","33.93 (min 31.25, max 37.14)" - "524288","3.76 (min 3.45, max 4.15)","28.78 (min 25.49, max 33.98)","11.28 (min 11.19, max 11.50)","33.96 (min 31.25, max 35.29)" - "1048576","3.77 (min 3.44, max 4.13)","28.45 (min 24.72, max 32.20)","11.18 (min 11.05, max 11.31)","37.67 (min 35.29, max 38.89)" - -.. rubric:: RAW - :name: am62lxx-evm-ospi-raw - -.. csv-table:: OSPI Raw Flash Driver - :header: "File size (Mbytes)","am62lxx_evm-fs: Raw Read Throughput (Mbytes/sec)" +| - "50","17.36 (min 16.95, max 17.67)" +EMMC +==== -EMMC Driver ------------ .. warning:: **IMPORTANT**: The performance numbers can be severely affected if the media is @@ -564,62 +553,51 @@ EMMC Driver re-mount in async mode. EMMC EXT4 FIO 1G -^^^^^^^^^^^^^^^^ +---------------- .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62lxx_evm-fs: Write EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Write EXT4 CPU Load (%)","am62lxx_evm-fs: Read EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Read EXT4 CPU Load (%)" - "1m","128.17 (min 128.00, max 129.00)","8.56 (min 8.30, max 8.86)","179.33 (min 179.00, max 180.00)","7.98 (min 7.62, max 8.22)" - "4m","127.50 (min 125.00, max 128.00)","5.46 (min 5.28, max 5.63)","179.50 (min 179.00, max 180.00)","4.95 (min 4.74, max 5.08)" - "4k","86.62 (min 86.40, max 86.90)","59.76 (min 59.27, max 60.23)","90.67 (min 88.60, max 93.50)","56.33 (min 55.48, max 56.62)" - "256k","122.83 (min 122.00, max 123.00)","11.83 (min 11.60, max 12.02)","178.33 (min 178.00, max 179.00)","10.08 (min 9.65, max 10.27)" - -EMMC RAW FIO 1G -^^^^^^^^^^^^^^^ - -.. csv-table:: EMMC RAW FIO 1G - :header: "Buffer size (bytes)","am62lxx_evm-fs: Write Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Write Raw CPU Load (%)","am62lxx_evm-fs: Read Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Read Raw CPU Load (%)" - - "1m","127.83 (min 122.00, max 129.00)","7.85 (min 7.48, max 7.99)","180.00","7.63 (min 7.16, max 8.07)" - "4m","129.00","5.33 (min 5.10, max 5.56)","180.00","4.67 (min 4.51, max 4.96)" - "4k","95.57 (min 93.50, max 96.40)","47.29 (min 45.85, max 48.11)","94.90 (min 94.00, max 95.80)","53.35 (min 52.55, max 54.08)" - "256k","122.67 (min 118.00, max 124.00)","9.76 (min 9.54, max 10.07)","178.17 (min 178.00, max 179.00)","9.31 (min 9.00, max 9.65)" + "1m","128.33 (min 128.00, max 129.00)","8.92 (min 8.60, max 9.20)","179.33 (min 179.00, max 180.00)","7.83 (min 7.36, max 8.35)" + "4m","128.67 (min 128.00, max 129.00)","5.52 (min 5.34, max 5.80)","179.33 (min 179.00, max 180.00)","5.08 (min 4.98, max 5.22)" + "4k","87.67 (min 86.30, max 88.60)","57.12 (min 56.47, max 57.84)","93.83 (min 90.80, max 97.50)","55.19 (min 54.00, max 57.17)" + "256k","123.00","12.30 (min 11.83, max 13.08)","177.67 (min 175.00, max 179.00)","10.40 (min 9.43, max 11.55)" EMMC EXT4 -^^^^^^^^^ +--------- .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62lxx_evm-fs: Write EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Write EXT4 CPU Load (%)","am62lxx_evm-fs: Read EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Read EXT4 CPU Load (%)" - "102400","90.91 (min 83.00, max 97.71)","20.38 (min 17.87, max 25.90)","180.36 (min 177.78, max 181.30)","37.46 (min 36.45, max 39.09)" - "262144","94.89 (min 83.93, max 100.22)","20.88 (min 18.72, max 25.51)","183.72 (min 177.40, max 185.29)","39.71 (min 36.70, max 50.88)" - "524288","95.62 (min 84.98, max 101.40)","20.84 (min 18.31, max 26.53)","186.81 (min 186.21, max 187.09)","32.93 (min 31.82, max 34.26)" - "1048576","95.42 (min 82.97, max 100.97)","21.14 (min 18.45, max 26.53)","186.89 (min 186.56, max 187.07)","32.73 (min 32.11, max 33.33)" - "5242880","96.51 (min 85.25, max 101.40)","20.64 (min 17.87, max 25.82)","186.37 (min 185.98, max 186.68)","32.98 (min 32.43, max 33.64)" + "102400","104.84 (min 95.48, max 108.41)","15.02 (min 13.33, max 20.00)","175.11 (min 158.41, max 183.74)","29.09 (min 25.19, max 33.64)" + "262144","106.77 (min 100.69, max 110.53)","12.13 (min 10.05, max 16.93)","185.91 (min 185.64, max 186.39)","27.61 (min 25.00, max 31.48)" + "524288","106.64 (min 99.47, max 110.24)","12.17 (min 9.74, max 15.57)","186.99 (min 186.68, max 187.16)","23.49 (min 20.91, max 25.00)" + "1048576","106.92 (min 102.07, max 109.98)","11.92 (min 10.26, max 15.71)","187.10 (min 187.05, max 187.17)","23.32 (min 22.22, max 24.11)" + "5242880","107.71 (min 100.42, max 111.86)","12.14 (min 10.47, max 15.05)","187.16 (min 187.07, max 187.23)","23.25 (min 22.73, max 24.07)" EMMC VFAT -^^^^^^^^^ +--------- .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62lxx_evm-fs: Write VFAT Throughput (Mbytes/sec)","am62lxx_evm-fs: Write VFAT CPU Load (%)","am62lxx_evm-fs: Read VFAT Throughput (Mbytes/sec)","am62lxx_evm-fs: Read VFAT CPU Load (%)" - "102400","32.73 (min 16.59, max 37.40)","15.81 (min 14.03, max 18.56)","173.62 (min 172.15, max 174.70)","39.10 (min 37.72, max 40.52)" - "262144","47.45 (min 20.52, max 63.09)","17.72 (min 13.82, max 23.74)","174.94 (min 164.63, max 177.64)","40.50 (min 36.61, max 55.74)" - "524288","56.02 (min 22.44, max 71.58)","18.76 (min 12.94, max 22.48)","177.43 (min 177.03, max 177.74)","32.31 (min 30.97, max 34.19)" - "1048576","61.76 (min 23.47, max 77.61)","19.71 (min 18.09, max 22.20)","177.54 (min 175.95, max 178.35)","34.28 (min 31.62, max 40.71)" - "5242880","71.24 (min 24.09, max 85.03)","20.62 (min 18.65, max 23.67)","176.90 (min 176.63, max 177.31)","32.18 (min 30.70, max 33.90)" + "102400","32.78 (min 16.04, max 37.19)","15.45 (min 12.96, max 20.27)","174.45 (min 173.98, max 175.01)","39.26 (min 36.52, max 43.97)" + "262144","47.10 (min 19.77, max 59.67)","18.04 (min 14.56, max 21.99)","175.94 (min 175.78, max 176.17)","36.81 (min 34.51, max 38.94)" + "524288","55.87 (min 21.16, max 65.96)","18.69 (min 15.92, max 25.61)","177.09 (min 176.63, max 177.56)","32.66 (min 31.03, max 34.19)" + "1048576","61.77 (min 22.07, max 76.62)","19.90 (min 18.12, max 22.93)","175.57 (min 174.58, max 177.21)","35.72 (min 33.04, max 40.52)" + "5242880","71.75 (min 23.10, max 85.16)","21.11 (min 19.44, max 22.40)","176.47 (min 176.20, max 176.65)","33.14 (min 32.48, max 34.45)" -UBoot EMMC Driver ------------------ +UBoot EMMC +---------- .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62lxx_evm-fs: Write Throughput (Kbytes/sec)","am62lxx_evm-fs: Read Throughput (Kbytes/sec)" - "2000000","112634.05 (min 110702.70, max 116198.58)","174765.56 (min 173375.66, max 175229.95)" - "4000000","125497.58 (min 122268.66, max 127254.37)","178086.96" + "2000000","111077.97","175229.95" + "4000000","124830.48","178086.96" MMCSD ------ +===== .. warning:: @@ -630,38 +608,27 @@ MMCSD re-mount in async mode. MMC EXT4 FIO 1G -^^^^^^^^^^^^^^^ +--------------- .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62lxx_evm-fs: Write EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Write EXT4 CPU Load (%)","am62lxx_evm-fs: Read EXT4 Throughput (Mbytes/sec)","am62lxx_evm-fs: Read EXT4 CPU Load (%)" - "1m","35.31 (min 18.70, max 42.30)","3.38 (min 2.18, max 4.24)","87.01 (min 86.40, max 87.50)","5.29 (min 4.42, max 5.60)" - "4m","34.96 (min 18.70, max 42.40)","2.36 (min 1.75, max 2.97)","86.47 (min 83.20, max 87.40)","3.02 (min 2.66, max 3.38)" - "4k","3.23 (min 2.74, max 4.29)","7.39 (min 5.78, max 9.52)","13.10 (min 12.90, max 13.50)","12.51 (min 11.54, max 13.83)" - "256k","31.90 (min 18.00, max 38.20)","5.09 (min 3.67, max 6.32)","83.10 (min 82.00, max 83.90)","6.51 (min 5.61, max 6.78)" - -MMC RAW FIO 1G -^^^^^^^^^^^^^^ - -.. csv-table:: MMC RAW FIO 1G - :header: "Buffer size (bytes)","am62lxx_evm-fs: Write Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Write Raw CPU Load (%)","am62lxx_evm-fs: Read Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Read Raw CPU Load (%)" - - "1m","31.02 (min 18.20, max 44.20)","2.80 (min 1.82, max 3.70)","87.63 (min 87.10, max 88.00)","4.61 (min 4.27, max 5.11)" - "4m","30.72 (min 18.20, max 43.60)","2.09 (min 1.32, max 2.68)","87.67 (min 87.00, max 88.10)","3.00 (min 2.61, max 3.48)" - "4k","3.45 (min 2.80, max 4.15)","5.24 (min 4.23, max 6.10)","13.32 (min 13.00, max 13.60)","11.41 (min 10.56, max 12.38)" - "256k","27.70 (min 17.60, max 38.60)","3.72 (min 2.58, max 4.57)","83.30 (min 82.10, max 84.10)","5.72 (min 4.84, max 6.15)" + "1m","42.60 (min 42.20, max 43.00)","4.13 (min 3.97, max 4.29)","86.40 (min 85.50, max 87.30)","5.47 (min 5.44, max 5.50)" + "4m","41.50 (min 40.30, max 42.70)","2.71 (min 2.63, max 2.78)","86.10 (min 84.70, max 87.50)","3.13 (min 3.12, max 3.14)" + "4k","2.86 (min 2.84, max 2.89)","7.03 (min 6.58, max 7.47)","12.95 (min 12.90, max 13.00)","12.01 (min 11.69, max 12.32)" + "256k","39.40 (min 38.90, max 39.90)","5.83 (min 5.38, max 6.28)","83.85 (min 83.70, max 84.00)","6.30 (min 6.26, max 6.34)" MMC EXT4 -^^^^^^^^ +-------- .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62lxx_evm-fs: Write Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Write Raw CPU Load (%)","am62lxx_evm-fs: Read Raw Throughput (Mbytes/sec)","am62lxx_evm-fs: Read Raw CPU Load (%)" - "102400","9.60 (min 8.52, max 10.57)","2.45 (min 1.99, max 3.24)","10.84 (min 10.64, max 11.16)","2.81 (min 2.45, max 3.08)" - "262144","9.45 (min 8.73, max 10.75)","2.46 (min 1.98, max 3.38)","10.99 (min 10.74, max 11.20)","2.35 (min 2.06, max 2.67)" - "524288","9.52 (min 8.75, max 10.56)","2.42 (min 1.92, max 3.28)","11.13 (min 10.90, max 11.49)","2.10 (min 1.94, max 2.18)" - "1048576","9.46 (min 8.86, max 10.52)","2.50 (min 1.96, max 3.77)","11.20 (min 11.08, max 11.38)","2.23 (min 2.01, max 2.66)" - "5242880","9.53 (min 8.79, max 10.77)","2.38 (min 1.91, max 3.95)","11.46 (min 11.05, max 12.02)","2.23 (min 1.96, max 2.49)" + "102400","10.74 (min 10.57, max 10.92)","1.53 (min 1.20, max 2.08)","10.92 (min 10.64, max 11.08)","2.10 (min 1.85, max 2.37)" + "262144","10.70 (min 10.25, max 11.01)","1.52 (min 1.11, max 2.42)","11.12 (min 10.99, max 11.19)","1.63 (min 1.34, max 1.83)" + "524288","10.74 (min 10.42, max 10.96)","1.33 (min 1.04, max 1.80)","11.36 (min 11.06, max 11.52)","1.34 (min 1.32, max 1.38)" + "1048576","10.83 (min 10.22, max 11.21)","1.42 (min 0.94, max 2.00)","11.50 (min 11.02, max 12.02)","1.36 (min 1.16, max 1.55)" + "5242880","10.93 (min 10.69, max 11.19)","1.29 (min 0.92, max 2.16)","12.01 (min 11.99, max 12.02)","1.49 (min 1.32, max 1.61)" The performance numbers were captured using the following: @@ -677,9 +644,9 @@ UBOOT MMCSD FAT .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62lxx_evm-fs: Write Throughput (Kbytes/sec)","am62lxx_evm-fs: Read Throughput (Kbytes/sec)" - "400000","17604.61 (min 12962.03, max 19692.31)","22714.58 (min 22505.49, max 22882.68)" - "800000","20272.80 (min 18004.40, max 21501.31)","23152.29 (min 23076.06, max 23206.80)" - "1000000","19450.02 (min 16015.64, max 21250.32)","23311.43 (min 23239.72, max 23339.03)" + "400000","28068.99 (min 10422.39, max 39009.52)","63923.25 (min 11636.36, max 83591.84)" + "800000","35679.64 (min 10908.12, max 49053.89)","67640.33 (min 11719.60, max 88086.02)" + "1000000","36338.73 (min 11033.00, max 47489.86)","69037.04 (min 11744.80, max 90021.98)" The performance numbers were captured using the following: @@ -687,89 +654,129 @@ The performance numbers were captured using the following: | -USB Driver ----------- +USB +=== USB Device Controller -^^^^^^^^^^^^^^^^^^^^^ +--------------------- .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT :header: "Number of Blocks","am62lxx_evm-fs: Throughput (MB/sec)" - "150","41.14 (min 36.20, max 43.50)" + "150","37.55 (min 36.00, max 40.60)" + .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT :header: "Number of Blocks","am62lxx_evm-fs: Throughput (MB/sec)" - "150","37.43 (min 31.00, max 41.40)" - -Low Power Performance ---------------------- - -Power Performance -^^^^^^^^^^^^^^^^^ - -These are power measurements taken while the device is in various low power modes like -Deep Sleep, RTC Only + DDR, and RTC Only. - -.. csv-table:: Deep Sleep Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core", "0.75", "4.06" - "soc_dvdd_1v8", "1.80", "1.48" - "soc_dvdd_3v3", "3.29", "3.59" - "vdda_1v8", "1.80", "1.38" - "vdd_lpddr4_pmic2", "1.10", "1.04" - "vdd_rtc", "0.74", "0.01" - "vdd_rtc_1v8", "1.80", "0.02" - "Total"," ","11.57" - -.. csv-table:: RTC Only + DDR Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core", "0.00", "0.00" - "soc_dvdd_1v8", "1.80", "1.41" - "soc_dvdd_3v3", "3.29", "1.09" - "vdda_1v8", "0.00", "0.00" - "vdd_lpddr4_pmic2", "1.10", "1.09" - "vdd_rtc", "0.75", "0.01" - "vdd_rtc_1v8", "1.80", "0.02" - "Total"," ","3.62" - -.. csv-table:: RTC Only Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core", "0.00", "0.00" - "soc_dvdd_1v8", "0.00", "0.00" - "soc_dvdd_3v3", "0.00", "0.00" - "vdda_1v8", "0.00", "0.00" - "vdd_lpddr4_pmic2", "0.00", "0.00" - "vdd_rtc", "0.75", "0.02" - "vdd_rtc_1v8", "1.80", "0.01" - "Total"," ","0.04" - -Resume Latency Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^ - -.. csv-table:: LPM Resume Latency Performance - :header: "Low Power Mode","Total Resume Latency (ms)" - - "RTC Only", "Full normal boot time (~seconds)" - "RTC Only + DDR", "456.44" - "Deep Sleep", "109.96" - -The performance numbers are measured without the Linux printk logs. To remove the -Linux printk logs, run the following commands in the terminal: - -.. code:: console - - # Detach kernel serial console - consoles=$(find /sys/bus/platform/devices/*.serial/ -name console) - for console in ${consoles}; do - echo -n N > ${console} - done + "150","31.63 (min 31.20, max 32.20)" +| -.. note:: +CRYPTO +====== + +OpenSSL Performance +------------------- + +.. csv-table:: OpenSSL Performance + :header: "Algorithm","Buffer Size (in bytes)","am62lxx_evm-fs: throughput (KBytes/Sec)" + + "aes-128-cbc","1024","820153.17 (min 819911.34, max 820533.93)" + "aes-128-cbc","16","88945.15 (min 88836.08, max 89002.73)" + "aes-128-cbc","16384","939063.98 (min 937519.79, max 940032.00)" + "aes-128-cbc","256","581491.09 (min 581083.82, max 581764.44)" + "aes-128-cbc","64","277695.28 (min 277243.09, max 278112.00)" + "aes-128-cbc","8192","930250.75 (min 929565.35, max 930660.35)" + "aes-128-ecb","1024","899457.71 (min 896871.42, max 901688.66)" + "aes-128-ecb","16","94404.48 (min 92735.88, max 95448.03)" + "aes-128-ecb","16384","1068426.58 (min 1067881.81, max 1068979.54)" + "aes-128-ecb","256","594218.41 (min 589874.01, max 597775.02)" + "aes-128-ecb","64","259665.47 (min 256530.28, max 261640.53)" + "aes-128-ecb","8192","1056195.93 (min 1055719.42, max 1057030.14)" + "aes-192-cbc","1024","635496.19 (min 634212.01, max 636192.09)" + "aes-192-cbc","16","84958.61 (min 83445.07, max 85605.02)" + "aes-192-cbc","16384","704898.39 (min 702851.75, max 706418.01)" + "aes-192-cbc","256","480254.10 (min 477063.59, max 481556.39)" + "aes-192-cbc","64","250222.52 (min 246972.93, max 251633.17)" + "aes-192-cbc","8192","701138.94 (min 700844.71, max 701322.58)" + "aes-192-ecb","1024","796659.71 (min 794786.82, max 797450.58)" + "aes-192-ecb","16","90855.18 (min 89027.61, max 91505.64)" + "aes-192-ecb","16384","929252.69 (min 928235.52, max 929824.77)" + "aes-192-ecb","256","544550.23 (min 540975.10, max 546148.27)" + "aes-192-ecb","64","245857.87 (min 242240.04, max 247096.23)" + "aes-192-ecb","8192","919371.09 (min 918964.91, max 919639.38)" + "aes-256-cbc","1024","533614.59 (min 532546.56, max 534222.85)" + "aes-256-cbc","16","82838.72 (min 81336.68, max 83377.88)" + "aes-256-cbc","16384","582748.84 (min 582489.43, max 582909.95)" + "aes-256-cbc","256","419748.05 (min 417439.23, max 420707.93)" + "aes-256-cbc","64","232853.02 (min 229933.29, max 233899.99)" + "aes-256-cbc","8192","579334.14 (min 579215.36, max 579499.35)" + "aes-256-ecb","1024","709590.02 (min 708103.51, max 710752.26)" + "aes-256-ecb","16","87171.66 (min 86051.96, max 88241.35)" + "aes-256-ecb","16384","813151.57 (min 812346.03, max 814159.19)" + "aes-256-ecb","256","488434.03 (min 485834.50, max 490525.18)" + "aes-256-ecb","64","232332.60 (min 230068.37, max 234072.00)" + "aes-256-ecb","8192","801607.00 (min 801270.44, max 802234.37)" + "des3","1024","4550.06 (min 4524.03, max 4562.26)" + "des3","16","4316.02 (min 4306.80, max 4319.40)" + "des3","16384","4564.31 (min 4560.21, max 4565.67)" + "des3","256","4548.33 (min 4545.19, max 4549.89)" + "des3","64","4497.16 (min 4490.35, max 4499.69)" + "des3","8192","4557.48 (min 4546.56, max 4565.67)" + "md5","1024","144407.38 (min 144327.00, max 144474.79)" + "md5","16","10784.95 (min 10761.82, max 10811.65)" + "md5","16384","179810.30 (min 179754.33, max 179869.01)" + "md5","256","88550.12 (min 88464.73, max 88638.55)" + "md5","64","34590.38 (min 34213.35, max 34737.00)" + "md5","8192","175732.74 (min 171794.43, max 177081.00)" + "sha1","1024","332384.51 (min 329384.28, max 333542.74)" + "sha1","16","12155.34 (min 12111.18, max 12188.26)" + "sha1","16384","557408.26 (min 555111.77, max 558410.41)" + "sha1","256","146047.32 (min 145553.07, max 146227.97)" + "sha1","64","44891.43 (min 44686.68, max 44967.42)" + "sha1","8192","534735.53 (min 533981.87, max 535374.51)" + "sha224","1024","319523.93 (min 317158.06, max 320510.98)" + "sha224","16","11875.43 (min 11692.84, max 11952.49)" + "sha224","16384","528467.29 (min 527777.79, max 529017.51)" + "sha224","256","141599.27 (min 139762.60, max 142371.58)" + "sha224","64","43810.79 (min 43110.55, max 44092.07)" + "sha224","8192","506508.63 (min 506161.83, max 507008.34)" + "sha256","1024","316993.96 (min 315484.84, max 318002.86)" + "sha256","16","11887.42 (min 11801.78, max 11934.95)" + "sha256","16384","528214.70 (min 527750.49, max 528438.61)" + "sha256","256","139981.29 (min 139145.39, max 140377.69)" + "sha256","64","43094.56 (min 42625.62, max 43298.41)" + "sha256","8192","505567.23 (min 504321.37, max 506585.09)" + "sha512","1024","99841.62 (min 99679.91, max 99963.90)" + "sha512","16","6977.87 (min 6945.17, max 7007.74)" + "sha512","16384","126693.38 (min 126621.01, max 126795.78)" + "sha512","256","59556.44 (min 59256.32, max 59719.51)" + "sha512","64","27937.12 (min 27885.93, max 28021.01)" + "sha512","8192","124555.95 (min 124523.86, max 124630.36)" + +.. csv-table:: OpenSSL CPU Load + :header: "Algorithm","am62lxx_evm-fs: CPU Load" + + "aes-128-cbc","98.75 (min 98.00, max 99.00)" + "aes-128-ecb","99.00" + "aes-192-cbc","99.00" + "aes-192-ecb","99.00" + "aes-256-cbc","99.00" + "aes-256-ecb","99.00" + "des3","99.00" + "md5","99.00" + "sha1","99.00" + "sha224","99.00" + "sha256","99.00" + "sha386","81.00 (min 78.00, max 85.00)" + "sha512","96.75 (min 90.00, max 99.00)" + +Listed for each algorithm are the code snippets used to run each +benchmark test. + +.. code-block:: console + + time -v openssl speed -elapsed -evp aes-128-cbc - The measurements shown are from using the default SDK with no extra optimizations. +| diff --git a/source/devices/AM62LX/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM62LX/linux/RT_Linux_Performance_Guide.rst index 74e78c36e..5f8a17a88 100644 --- a/source/devices/AM62LX/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM62LX/linux/RT_Linux_Performance_Guide.rst @@ -1,8 +1,10 @@ -======================================= - RT-linux 11.02.08.02 Performance Guide -======================================= +################################### +RT-linux 12.00.00 Performance Guide +################################### -.. rubric:: **Read This First** +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -15,9 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -| - -.. rubric:: About This Manual +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Process SDK Linux package. This document should be @@ -25,60 +27,13 @@ used in conjunction with release notes and user guides provided with the Process SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ | +***************** System Benchmarks ------------------ - -Stress-ng and Cyclic Test - -.. _RT-linux-performance: - -stress-ng (next-generation) will stress test a embedded platform in various selectable ways. -It was designed to exercise various physical subsystems as well as the various -operating system kernel interfaces. stress-ng can also measure test throughput rates; -this can be useful to observe performance changes across different operating system or types of hardware. - -Cyclictest is most commonly used for benchmarking RT systems. -It is one of the most frequently used tools for evaluating the relative performance of real-time systems. -Some performance tests which use Cyclictest are System benchmarking, Latency debugging with tracing and -approximating application performance. - -Test commands used for running stress-ng and cyclictest together - -.. code:: console +***************** - stress-ng --cpu-method=all -c 4 & - cyclictest -m -Sp80 -D6h -h400 -i200 -M -q - -The following summarizes the latencies observed using the yocto based -default SDK image: - -.. note:: - - Using the OP-TEE TRNG driver can impact this benchmark's performance due to - frequent context switching between Normal World (Linux) and Secure World (OP-TEE), - that occurs when the kernel's hardware random number generator interface - (hwrng) requests entropy from the secure TRNG to replenish the Linux entropy - pool. - - The Linux TRNG driver can mitigate these latency spikes. This involves - enabling the Pseudo RNG driver in OP-TEE as documented in the Foundational - Components section: :ref:`building-optee-with-prng`, and enabling the RNG - node in the Linux kernel device tree. This way the HW TRNG is accessed from - the kernel itself. - - -.. csv-table:: - :header: "Latencies","CPU 0","CPU 1" - - "Minimum (usec)","5","5" - "Average (usec)","8","7" - "Maximum (usec)","62","47" - -.. image:: img/rt-cpu-method-all-latency-histogram.png +| diff --git a/source/devices/AM62PX/linux/Linux_Performance_Guide.rst b/source/devices/AM62PX/linux/Linux_Performance_Guide.rst index 300f56a3a..a7ee46727 100644 --- a/source/devices/AM62PX/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62PX/linux/Linux_Performance_Guide.rst @@ -1,10 +1,10 @@ +################################# + Linux 12.00.00 Performance Guide +################################# -=================================== -Linux 11.02.08.02 Performance Guide -=================================== - -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -17,8 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -26,17 +27,20 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ +| + +***************** System Benchmarks ------------------ +***************** + +| LMBench -^^^^^^^ +======= + LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at @@ -63,124 +67,124 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "af_unix_sock_stream_latency (microsec)","30.38 (min 29.49, max 31.11)" - "af_unix_socket_stream_bandwidth (mb\s)","1141.99 (min 1114.62, max 1178.85)" - "bw_file_rd-io-1mb (mb/s)","1420.50 (min 1397.38, max 1467.89)" - "bw_file_rd-o2c-1mb (mb/s)","756.14 (min 707.71, max 787.56)" - "bw_mem-bcopy-16mb (mb/s)","1893.33 (min 1820.46, max 1973.12)" - "bw_mem-bcopy-1mb (mb/s)","2039.91 (min 1948.56, max 2126.53)" - "bw_mem-bcopy-2mb (mb/s)","1763.70 (min 1689.19, max 1838.80)" - "bw_mem-bcopy-4mb (mb/s)","1778.31 (min 1643.39, max 1932.06)" - "bw_mem-bcopy-8mb (mb/s)","1862.34 (min 1802.61, max 1939.39)" - "bw_mem-bzero-16mb (mb/s)","7941.37 (min 7807.42, max 8116.33)" - "bw_mem-bzero-1mb (mb/s)","4983.96 (min 1948.56, max 8117.51)" - "bw_mem-bzero-2mb (mb/s)","4846.65 (min 1689.19, max 8105.73)" - "bw_mem-bzero-4mb (mb/s)","4858.50 (min 1643.39, max 8110.85)" - "bw_mem-bzero-8mb (mb/s)","4901.14 (min 1802.61, max 8110.85)" - "bw_mem-cp-16mb (mb/s)","936.73 (min 894.15, max 985.16)" - "bw_mem-cp-1mb (mb/s)","4594.51 (min 890.31, max 8403.36)" - "bw_mem-cp-2mb (mb/s)","4486.54 (min 865.05, max 8255.56)" - "bw_mem-cp-4mb (mb/s)","4485.96 (min 926.14, max 8185.54)" - "bw_mem-cp-8mb (mb/s)","4481.34 (min 948.77, max 8154.94)" - "bw_mem-fcp-16mb (mb/s)","1764.54 (min 1699.96, max 1837.60)" - "bw_mem-fcp-1mb (mb/s)","4843.70 (min 1689.51, max 8117.51)" - "bw_mem-fcp-2mb (mb/s)","4796.57 (min 1575.05, max 8105.73)" - "bw_mem-fcp-4mb (mb/s)","4826.26 (min 1609.87, max 8110.85)" - "bw_mem-fcp-8mb (mb/s)","4831.18 (min 1631.82, max 8110.85)" - "bw_mem-frd-16mb (mb/s)","1908.07 (min 1837.18, max 2011.06)" - "bw_mem-frd-1mb (mb/s)","1907.50 (min 1689.51, max 2142.09)" - "bw_mem-frd-2mb (mb/s)","1705.84 (min 1575.05, max 1836.88)" - "bw_mem-frd-4mb (mb/s)","1764.63 (min 1609.87, max 1909.31)" - "bw_mem-frd-8mb (mb/s)","1789.25 (min 1631.82, max 1998.75)" - "bw_mem-fwr-16mb (mb/s)","7958.93 (min 7822.69, max 8134.21)" - "bw_mem-fwr-1mb (mb/s)","5153.95 (min 1974.16, max 8403.36)" - "bw_mem-fwr-2mb (mb/s)","4904.86 (min 1597.44, max 8255.56)" - "bw_mem-fwr-4mb (mb/s)","4913.69 (min 1659.06, max 8185.54)" - "bw_mem-fwr-8mb (mb/s)","4917.48 (min 1674.52, max 8154.94)" - "bw_mem-rd-16mb (mb/s)","1983.53 (min 1918.24, max 2065.32)" - "bw_mem-rd-1mb (mb/s)","2035.54 (min 1708.82, max 2331.46)" - "bw_mem-rd-2mb (mb/s)","1794.96 (min 1510.00, max 2037.35)" - "bw_mem-rd-4mb (mb/s)","1864.28 (min 1662.05, max 2055.85)" - "bw_mem-rd-8mb (mb/s)","1901.83 (min 1763.28, max 2069.05)" - "bw_mem-rdwr-16mb (mb/s)","1853.51 (min 1684.21, max 1942.45)" - "bw_mem-rdwr-1mb (mb/s)","1329.35 (min 890.31, max 1855.29)" - "bw_mem-rdwr-2mb (mb/s)","1231.50 (min 865.05, max 1661.96)" - "bw_mem-rdwr-4mb (mb/s)","1342.24 (min 926.14, max 1834.58)" - "bw_mem-rdwr-8mb (mb/s)","1405.93 (min 948.77, max 1942.22)" - "bw_mem-wr-16mb (mb/s)","1883.72 (min 1802.41, max 1975.55)" - "bw_mem-wr-1mb (mb/s)","1779.17 (min 1614.06, max 2043.60)" - "bw_mem-wr-2mb (mb/s)","1589.31 (min 1490.59, max 1853.57)" - "bw_mem-wr-4mb (mb/s)","1739.42 (min 1631.10, max 1869.45)" - "bw_mem-wr-8mb (mb/s)","1824.68 (min 1672.24, max 1942.22)" - "bw_mmap_rd-mo-1mb (mb/s)","2177.27 (min 2111.56, max 2271.54)" - "bw_mmap_rd-o2c-1mb (mb/s)","772.38 (min 680.85, max 826.72)" - "bw_pipe (mb/s)","783.15 (min 729.29, max 819.18)" - "bw_unix (mb/s)","1141.99 (min 1114.62, max 1178.85)" - "lat_connect (us)","56.81 (min 56.67, max 57.00)" - "lat_ctx-2-128k (us)","7.83 (min 7.60, max 8.13)" - "lat_ctx-2-256k (us)","7.72 (min 6.67, max 11.90)" - "lat_ctx-4-128k (us)","7.42 (min 6.78, max 7.96)" - "lat_ctx-4-256k (us)","6.99 (min 6.18, max 8.04)" - "lat_fs-0k (num_files)","239.71 (min 221.00, max 253.00)" - "lat_fs-10k (num_files)","117.29 (min 106.00, max 136.00)" - "lat_fs-1k (num_files)","160.43 (min 145.00, max 179.00)" - "lat_fs-4k (num_files)","159.43 (min 142.00, max 176.00)" - "lat_mem_rd-stride128-sz1000k (ns)","30.55 (min 29.36, max 31.41)" - "lat_mem_rd-stride128-sz125k (ns)","5.57 (min 5.54, max 5.64)" - "lat_mem_rd-stride128-sz250k (ns)","5.84 (min 5.83, max 5.85)" - "lat_mem_rd-stride128-sz31k (ns)","3.15 (min 2.16, max 4.66)" - "lat_mem_rd-stride128-sz50 (ns)","2.15" - "lat_mem_rd-stride128-sz500k (ns)","10.88 (min 9.02, max 12.76)" - "lat_mem_rd-stride128-sz62k (ns)","5.24 (min 5.20, max 5.26)" - "lat_mmap-1m (us)","52.57 (min 50.00, max 57.00)" - "lat_ops-double-add (ns)","2.86" - "lat_ops-double-div (ns)","15.74 (min 15.74, max 15.75)" - "lat_ops-double-mul (ns)","2.86" - "lat_ops-float-add (ns)","2.86" - "lat_ops-float-div (ns)","9.30 (min 9.30, max 9.31)" - "lat_ops-float-mul (ns)","2.86" - "lat_ops-int-add (ns)","0.72" - "lat_ops-int-bit (ns)","0.48" - "lat_ops-int-div (ns)","4.29" - "lat_ops-int-mod (ns)","4.53 (min 4.53, max 4.54)" - "lat_ops-int-mul (ns)","3.08 (min 3.07, max 3.09)" - "lat_ops-int64-add (ns)","0.72" - "lat_ops-int64-bit (ns)","0.48" - "lat_ops-int64-div (ns)","6.80" - "lat_ops-int64-mod (ns)","5.25" - "lat_ops-int64-mul (ns)","3.56 (min 3.55, max 3.57)" - "lat_pagefault (us)","0.52 (min 0.50, max 0.54)" - "lat_pipe (us)","25.74 (min 25.17, max 26.44)" - "lat_proc-exec (us)","709.41 (min 689.88, max 728.75)" - "lat_proc-fork (us)","616.86 (min 602.89, max 633.11)" + "af_unix_sock_stream_latency (microsec)","27.25 (min 23.37, max 31.39)" + "af_unix_socket_stream_bandwidth (mb\s)","1098.87 (min 1064.86, max 1145.16)" + "bw_file_rd-io-1mb (mb/s)","1297.39 (min 1250.45, max 1353.70)" + "bw_file_rd-o2c-1mb (mb/s)","698.21 (min 626.27, max 743.36)" + "bw_mem-bcopy-16mb (mb/s)","1819.35 (min 1787.11, max 1858.30)" + "bw_mem-bcopy-1mb (mb/s)","1968.82 (min 1926.11, max 2020.20)" + "bw_mem-bcopy-2mb (mb/s)","1636.56 (min 1562.99, max 1757.47)" + "bw_mem-bcopy-4mb (mb/s)","1760.75 (min 1715.76, max 1810.50)" + "bw_mem-bcopy-8mb (mb/s)","1783.97 (min 1721.36, max 1831.92)" + "bw_mem-bzero-16mb (mb/s)","7382.69 (min 7082.78, max 7846.98)" + "bw_mem-bzero-1mb (mb/s)","4669.86 (min 1926.11, max 7834.89)" + "bw_mem-bzero-2mb (mb/s)","4501.42 (min 1562.99, max 7834.76)" + "bw_mem-bzero-4mb (mb/s)","4568.55 (min 1715.76, max 7836.15)" + "bw_mem-bzero-8mb (mb/s)","4584.74 (min 1721.36, max 7857.26)" + "bw_mem-cp-16mb (mb/s)","900.30 (min 885.79, max 917.43)" + "bw_mem-cp-1mb (mb/s)","4257.87 (min 872.75, max 8091.21)" + "bw_mem-cp-2mb (mb/s)","4187.65 (min 858.86, max 7981.14)" + "bw_mem-cp-4mb (mb/s)","4183.31 (min 906.00, max 7906.56)" + "bw_mem-cp-8mb (mb/s)","4182.61 (min 932.73, max 7880.48)" + "bw_mem-fcp-16mb (mb/s)","1699.60 (min 1673.12, max 1742.16)" + "bw_mem-fcp-1mb (mb/s)","4518.12 (min 1637.85, max 7834.89)" + "bw_mem-fcp-2mb (mb/s)","4484.17 (min 1506.70, max 7834.76)" + "bw_mem-fcp-4mb (mb/s)","4516.30 (min 1596.17, max 7836.15)" + "bw_mem-fcp-8mb (mb/s)","4531.23 (min 1649.14, max 7857.26)" + "bw_mem-frd-16mb (mb/s)","1782.34 (min 1669.45, max 1859.82)" + "bw_mem-frd-1mb (mb/s)","1757.90 (min 1637.85, max 1979.84)" + "bw_mem-frd-2mb (mb/s)","1652.86 (min 1506.70, max 1781.58)" + "bw_mem-frd-4mb (mb/s)","1699.36 (min 1596.17, max 1833.74)" + "bw_mem-frd-8mb (mb/s)","1717.18 (min 1624.20, max 1853.14)" + "bw_mem-fwr-16mb (mb/s)","7398.31 (min 7099.54, max 7855.97)" + "bw_mem-fwr-1mb (mb/s)","4739.74 (min 1767.92, max 8091.21)" + "bw_mem-fwr-2mb (mb/s)","4599.47 (min 1652.07, max 7981.14)" + "bw_mem-fwr-4mb (mb/s)","4594.96 (min 1597.02, max 7906.56)" + "bw_mem-fwr-8mb (mb/s)","4588.07 (min 1624.20, max 7880.48)" + "bw_mem-rd-16mb (mb/s)","1879.38 (min 1823.36, max 1934.00)" + "bw_mem-rd-1mb (mb/s)","1948.25 (min 1650.77, max 2211.57)" + "bw_mem-rd-2mb (mb/s)","1700.50 (min 1484.56, max 1916.32)" + "bw_mem-rd-4mb (mb/s)","1771.38 (min 1632.43, max 1932.99)" + "bw_mem-rd-8mb (mb/s)","1808.54 (min 1708.12, max 1918.93)" + "bw_mem-rdwr-16mb (mb/s)","1812.20 (min 1768.15, max 1847.36)" + "bw_mem-rdwr-1mb (mb/s)","1279.41 (min 872.75, max 1768.97)" + "bw_mem-rdwr-2mb (mb/s)","1176.63 (min 858.86, max 1505.57)" + "bw_mem-rdwr-4mb (mb/s)","1258.98 (min 906.00, max 1676.45)" + "bw_mem-rdwr-8mb (mb/s)","1349.98 (min 932.73, max 1789.91)" + "bw_mem-wr-16mb (mb/s)","1789.42 (min 1731.98, max 1835.07)" + "bw_mem-wr-1mb (mb/s)","1704.44 (min 1634.58, max 1887.44)" + "bw_mem-wr-2mb (mb/s)","1505.81 (min 1451.12, max 1630.99)" + "bw_mem-wr-4mb (mb/s)","1634.87 (min 1510.95, max 1731.85)" + "bw_mem-wr-8mb (mb/s)","1748.27 (min 1703.76, max 1789.91)" + "bw_mmap_rd-mo-1mb (mb/s)","2006.64 (min 1947.10, max 2076.84)" + "bw_mmap_rd-o2c-1mb (mb/s)","705.96 (min 677.62, max 734.75)" + "bw_pipe (mb/s)","754.85 (min 721.80, max 814.87)" + "bw_unix (mb/s)","1098.87 (min 1064.86, max 1145.16)" + "lat_connect (us)","68.66 (min 51.27, max 81.66)" + "lat_ctx-2-128k (us)","9.35 (min 8.02, max 10.68)" + "lat_ctx-2-256k (us)","24.48 (min 8.17, max 45.71)" + "lat_ctx-4-128k (us)","17.09 (min 8.06, max 30.58)" + "lat_ctx-4-256k (us)","43.70 (min 7.31, max 102.29)" + "lat_fs-0k (num_files)","259.00 (min 207.00, max 308.00)" + "lat_fs-10k (num_files)","121.40 (min 101.00, max 154.00)" + "lat_fs-1k (num_files)","175.80 (min 163.00, max 196.00)" + "lat_fs-4k (num_files)","168.20 (min 150.00, max 183.00)" + "lat_mem_rd-stride128-sz1000k (ns)","31.90 (min 31.10, max 32.39)" + "lat_mem_rd-stride128-sz125k (ns)","5.96 (min 5.53, max 6.23)" + "lat_mem_rd-stride128-sz250k (ns)","6.26 (min 5.83, max 6.54)" + "lat_mem_rd-stride128-sz31k (ns)","2.72 (min 2.16, max 4.19)" + "lat_mem_rd-stride128-sz50 (ns)","2.30 (min 2.15, max 2.40)" + "lat_mem_rd-stride128-sz500k (ns)","12.15 (min 11.18, max 12.90)" + "lat_mem_rd-stride128-sz62k (ns)","5.61 (min 5.23, max 5.88)" + "lat_mmap-1m (us)","57.20 (min 56.00, max 62.00)" + "lat_ops-double-add (ns)","3.07 (min 2.86, max 3.21)" + "lat_ops-double-div (ns)","16.88 (min 15.75, max 17.64)" + "lat_ops-double-mul (ns)","3.07 (min 2.86, max 3.21)" + "lat_ops-float-add (ns)","3.07 (min 2.86, max 3.21)" + "lat_ops-float-div (ns)","10.02 (min 9.30, max 10.64)" + "lat_ops-float-mul (ns)","3.07 (min 2.86, max 3.22)" + "lat_ops-int-add (ns)","0.77 (min 0.72, max 0.80)" + "lat_ops-int-bit (ns)","0.51 (min 0.48, max 0.53)" + "lat_ops-int-div (ns)","4.60 (min 4.29, max 4.81)" + "lat_ops-int-mod (ns)","4.86 (min 4.53, max 5.08)" + "lat_ops-int-mul (ns)","3.33 (min 3.05, max 3.49)" + "lat_ops-int64-add (ns)","0.77 (min 0.72, max 0.80)" + "lat_ops-int64-bit (ns)","0.51 (min 0.48, max 0.53)" + "lat_ops-int64-div (ns)","7.29 (min 6.80, max 7.62)" + "lat_ops-int64-mod (ns)","5.63 (min 5.25, max 5.88)" + "lat_ops-int64-mul (ns)","3.82 (min 3.54, max 4.07)" + "lat_pagefault (us)","0.47 (min 0.45, max 0.48)" + "lat_pipe (us)","22.97 (min 21.04, max 24.59)" + "lat_proc-exec (us)","746.80 (min 706.88, max 781.00)" + "lat_proc-fork (us)","652.35 (min 620.22, max 681.50)" "lat_proc-proccall (us)","0.01" - "lat_select (us)","33.97 (min 33.79, max 34.06)" - "lat_sem (us)","3.04 (min 2.65, max 3.79)" - "lat_sig-catch (us)","5.54 (min 5.27, max 5.77)" - "lat_sig-install (us)","0.67 (min 0.65, max 0.71)" - "lat_sig-prot (us)","0.65 (min 0.50, max 0.76)" - "lat_syscall-fstat (us)","1.99 (min 1.90, max 2.07)" - "lat_syscall-null (us)","0.46 (min 0.46, max 0.50)" - "lat_syscall-open (us)","188.47 (min 153.56, max 251.41)" - "lat_syscall-read (us)","0.83 (min 0.80, max 0.88)" - "lat_syscall-stat (us)","4.79 (min 4.67, max 4.97)" - "lat_syscall-write (us)","0.78 (min 0.75, max 0.84)" - "lat_tcp (us)","0.92 (min 0.91, max 0.97)" - "lat_unix (us)","30.38 (min 29.49, max 31.11)" - "latency_for_0.50_mb_block_size (nanosec)","10.88 (min 9.02, max 12.76)" - "latency_for_1.00_mb_block_size (nanosec)","15.28 (min 0.00, max 31.41)" - "pipe_bandwidth (mb\s)","783.15 (min 729.29, max 819.18)" - "pipe_latency (microsec)","25.74 (min 25.17, max 26.44)" + "lat_select (us)","33.85 (min 31.32, max 35.89)" + "lat_sem (us)","2.83 (min 2.58, max 3.13)" + "lat_sig-catch (us)","6.03 (min 5.51, max 6.37)" + "lat_sig-install (us)","0.69 (min 0.58, max 0.80)" + "lat_sig-prot (us)","0.90 (min 0.80, max 0.97)" + "lat_syscall-fstat (us)","1.93 (min 1.74, max 2.10)" + "lat_syscall-null (us)","0.45 (min 0.37, max 0.56)" + "lat_syscall-open (us)","222.92 (min 145.19, max 448.21)" + "lat_syscall-read (us)","0.80 (min 0.74, max 0.86)" + "lat_syscall-stat (us)","4.42 (min 3.96, max 4.79)" + "lat_syscall-write (us)","0.75 (min 0.67, max 0.82)" + "lat_tcp (us)","0.92 (min 0.76, max 1.14)" + "lat_unix (us)","27.25 (min 23.37, max 31.39)" + "latency_for_0.50_mb_block_size (nanosec)","12.15 (min 11.18, max 12.90)" + "latency_for_1.00_mb_block_size (nanosec)","15.95 (min 0.00, max 32.39)" + "pipe_bandwidth (mb\s)","754.85 (min 721.80, max 814.87)" + "pipe_latency (microsec)","22.97 (min 21.04, max 24.59)" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","33.97 (min 33.79, max 34.06)" - "semaphore_latency (microsec)","3.04 (min 2.65, max 3.79)" - "signal_handler_latency (microsec)","0.67 (min 0.65, max 0.71)" - "signal_handler_overhead (microsec)","5.54 (min 5.27, max 5.77)" - "tcp_ip_connection_cost_to_localhost (microsec)","56.81 (min 56.67, max 57.00)" - "tcp_latency_using_localhost (microsec)","0.92 (min 0.91, max 0.97)" + "select_on_200_tcp_fds (microsec)","33.85 (min 31.32, max 35.89)" + "semaphore_latency (microsec)","2.83 (min 2.58, max 3.13)" + "signal_handler_latency (microsec)","0.69 (min 0.58, max 0.80)" + "signal_handler_overhead (microsec)","6.03 (min 5.51, max 6.37)" + "tcp_ip_connection_cost_to_localhost (microsec)","68.66 (min 51.27, max 81.66)" + "tcp_latency_using_localhost (microsec)","0.92 (min 0.76, max 1.14)" Dhrystone -^^^^^^^^^ +========= Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. @@ -197,12 +201,13 @@ Execute the benchmark with the following: .. csv-table:: Dhrystone Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "cpu_clock (mhz)","1400.00" - "dhrystone_per_mhz (dmips/mhz)","2.91 (min 2.90, max 3.00)" - "dhrystone_per_second (dhrystonep)","7180649.93 (min 7142857.00, max 7407407.50)" + "cpu_clock (mhz)","1310.00 (min 1250.00, max 1400.00)" + "dhrystone_per_mhz (dmips/mhz)","2.76 (min 2.70, max 2.80)" + "dhrystone_per_second (dhrystonep)","6323683.00 (min 5882353.00, max 6896551.50)" Whetstone -^^^^^^^^^ +========= + Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -214,20 +219,22 @@ Execute the benchmark with the following: .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "whetstone (mips)","6428.57 (min 5000.00, max 10000.00)" + "whetstone (mips)","5000.00" Linpack -^^^^^^^ +======= + Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "linpack (kflops)","575572.00 (min 571460.00, max 577031.00)" + "linpack (kflops)","559402.33 (min 518395.00, max 580053.00)" NBench -^^^^^^ +====== + NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at @@ -237,18 +244,19 @@ https://nbench.io/articles/index.html .. csv-table:: NBench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "assignment (iterations)","14.48 (min 14.45, max 14.50)" - "fourier (iterations)","22831.29 (min 22831.00, max 22832.00)" - "fp_emulation (iterations)","215.64 (min 215.64, max 215.65)" - "huffman (iterations)","1184.13 (min 1183.30, max 1184.40)" - "idea (iterations)","3444.83 (min 3444.70, max 3444.90)" - "lu_decomposition (iterations)","529.08 (min 526.35, max 533.03)" - "neural_net (iterations)","8.65 (min 8.65, max 8.66)" - "numeric_sort (iterations)","625.73 (min 617.87, max 629.99)" - "string_sort (iterations)","163.94 (min 163.93, max 163.94)" + "assignment (iterations)","13.25 (min 12.62, max 14.16)" + "fourier (iterations)","19455.60 (min 18561.00, max 20791.00)" + "fp_emulation (iterations)","191.26 (min 182.55, max 204.44)" + "huffman (iterations)","1111.14 (min 1063.00, max 1190.80)" + "idea (iterations)","2869.98 (min 2738.50, max 3067.40)" + "lu_decomposition (iterations)","496.24 (min 469.90, max 532.70)" + "neural_net (iterations)","9.03 (min 8.64, max 9.68)" + "numeric_sort (iterations)","506.05 (min 479.67, max 542.75)" + "string_sort (iterations)","157.42 (min 150.17, max 168.31)" Stream -^^^^^^ +====== + STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -268,13 +276,14 @@ Execute the benchmark with the following: .. csv-table:: Stream Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "add (mb/s)","2825.76 (min 2698.10, max 2879.00)" - "copy (mb/s)","3941.66 (min 3718.70, max 4027.40)" - "scale (mb/s)","3624.44 (min 3481.60, max 3704.10)" - "triad (mb/s)","2492.96 (min 2403.50, max 2530.10)" + "add (mb/s)","2595.38 (min 2519.20, max 2715.50)" + "copy (mb/s)","3692.72 (min 3632.10, max 3761.80)" + "scale (mb/s)","3369.20 (min 3233.70, max 3519.80)" + "triad (mb/s)","2328.48 (min 2274.80, max 2413.50)" CoreMarkPro -^^^^^^^^^^^ +=========== + CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, @@ -284,33 +293,34 @@ and floating-point workloads, and data sets for utilizing larger memory subsyste .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "cjpeg-rose7-preset (workloads/)","41.97 (min 41.84, max 42.02)" - "core (workloads/)","0.30" - "coremark-pro ()","927.45 (min 906.93, max 947.60)" - "linear_alg-mid-100x100-sp (workloads/)","14.68 (min 14.68, max 14.70)" - "loops-all-mid-10k-sp (workloads/)","0.71 (min 0.71, max 0.72)" - "nnet_test (workloads/)","1.09" - "parser-125k (workloads/)","8.83 (min 8.77, max 8.85)" - "radix2-big-64k (workloads/)","62.81 (min 53.18, max 74.52)" - "sha-test (workloads/)","81.68 (min 81.30, max 81.97)" - "zip-test (workloads/)","22.15 (min 21.74, max 22.22)" + "cjpeg-rose7-preset (workloads/)","39.26 (min 37.31, max 42.19)" + "core (workloads/)","0.28 (min 0.27, max 0.30)" + "coremark-pro ()","878.42 (min 816.27, max 956.64)" + "linear_alg-mid-100x100-sp (workloads/)","13.73 (min 13.10, max 14.68)" + "loops-all-mid-10k-sp (workloads/)","0.67 (min 0.64, max 0.71)" + "nnet_test (workloads/)","1.01 (min 0.96, max 1.08)" + "parser-125k (workloads/)","8.48 (min 7.94, max 9.35)" + "radix2-big-64k (workloads/)","63.41 (min 48.48, max 75.74)" + "sha-test (workloads/)","76.21 (min 72.46, max 81.30)" + "zip-test (workloads/)","21.17 (min 20.00, max 23.26)" .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "cjpeg-rose7-preset (workloads/)","83.45 (min 82.64, max 84.03)" - "core (workloads/)","0.60" - "coremark-pro ()","1670.65 (min 1656.51, max 1691.47)" - "linear_alg-mid-100x100-sp (workloads/)","29.34 (min 29.31, max 29.36)" - "loops-all-mid-10k-sp (workloads/)","1.31 (min 1.30, max 1.31)" - "nnet_test (workloads/)","2.17 (min 2.16, max 2.17)" - "parser-125k (workloads/)","13.95 (min 13.51, max 14.49)" - "radix2-big-64k (workloads/)","71.62 (min 68.31, max 76.15)" - "sha-test (workloads/)","162.61 (min 161.29, max 163.93)" - "zip-test (workloads/)","42.55" + "cjpeg-rose7-preset (workloads/)","78.14 (min 74.07, max 84.03)" + "core (workloads/)","0.56 (min 0.54, max 0.60)" + "coremark-pro ()","1582.94 (min 1487.40, max 1713.31)" + "linear_alg-mid-100x100-sp (workloads/)","27.45 (min 26.19, max 29.36)" + "loops-all-mid-10k-sp (workloads/)","1.24 (min 1.18, max 1.32)" + "nnet_test (workloads/)","2.01 (min 1.92, max 2.15)" + "parser-125k (workloads/)","14.04 (min 12.05, max 16.67)" + "radix2-big-64k (workloads/)","66.02 (min 63.30, max 69.22)" + "sha-test (workloads/)","151.47 (min 144.93, max 161.29)" + "zip-test (workloads/)","41.87 (min 38.46, max 46.51)" MultiBench -^^^^^^^^^^ +========== + MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -329,51 +339,53 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "4m-check (workloads/)","419.94 (min 412.34, max 431.18)" - "4m-check-reassembly (workloads/)","123.93 (min 120.34, max 128.54)" - "4m-check-reassembly-tcp (workloads/)","59.99 (min 59.10, max 61.27)" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","33.80 (min 33.24, max 34.68)" - "4m-check-reassembly-tcp-x264w2 (workloads/)","1.89 (min 1.87, max 1.91)" - "4m-cmykw2 (workloads/)","243.33 (min 238.66, max 246.61)" - "4m-cmykw2-rotatew2 (workloads/)","50.36 (min 49.71, max 51.37)" - "4m-reassembly (workloads/)","83.96 (min 82.30, max 86.28)" - "4m-rotatew2 (workloads/)","53.30 (min 52.63, max 54.26)" - "4m-tcp-mixed (workloads/)","118.90 (min 118.52, max 119.40)" - "4m-x264w2 (workloads/)","1.95 (min 1.80, max 2.00)" - "idct-4m (workloads/)","19.21 (min 19.18, max 19.23)" - "idct-4mw1 (workloads/)","19.21 (min 19.18, max 19.23)" - "ippktcheck-4m (workloads/)","421.59 (min 413.22, max 431.63)" - "ippktcheck-4mw1 (workloads/)","421.60 (min 412.61, max 431.33)" - "ipres-4m (workloads/)","113.14 (min 110.54, max 116.73)" - "ipres-4mw1 (workloads/)","113.31 (min 110.13, max 116.64)" - "md5-4m (workloads/)","27.53 (min 27.10, max 27.88)" - "md5-4mw1 (workloads/)","27.73 (min 27.45, max 27.98)" - "rgbcmyk-4m (workloads/)","64.48 (min 62.70, max 65.81)" - "rgbcmyk-4mw1 (workloads/)","64.46 (min 62.68, max 65.77)" - "rotate-4ms1 (workloads/)","23.74 (min 23.38, max 24.15)" - "rotate-4ms1w1 (workloads/)","23.76 (min 23.41, max 24.21)" - "rotate-4ms64 (workloads/)","24.00 (min 23.66, max 24.49)" - "rotate-4ms64w1 (workloads/)","24.34 (min 23.64, max 26.53)" - "x264-4mq (workloads/)","0.58 (min 0.57, max 0.58)" - "x264-4mqw1 (workloads/)","0.57 (min 0.53, max 0.58)" + "4m-check (workloads/)","400.01 (min 389.29, max 419.89)" + "4m-check-reassembly (workloads/)","113.31 (min 110.38, max 118.06)" + "4m-check-reassembly-tcp (workloads/)","57.65 (min 55.93, max 60.83)" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","32.13 (min 30.91, max 34.38)" + "4m-check-reassembly-tcp-x264w2 (workloads/)","1.75 (min 1.66, max 1.90)" + "4m-cmykw2 (workloads/)","230.97 (min 222.72, max 247.22)" + "4m-cmykw2-rotatew2 (workloads/)","44.27 (min 42.64, max 46.69)" + "4m-reassembly (workloads/)","75.98 (min 74.02, max 79.30)" + "4m-rotatew2 (workloads/)","49.28 (min 47.37, max 52.49)" + "4m-tcp-mixed (workloads/)","119.41 (min 115.11, max 128.00)" + "4m-x264w2 (workloads/)","1.75 (min 1.75, max 1.76)" + "idct-4m (workloads/)","17.85 (min 17.17, max 19.20)" + "idct-4mw1 (workloads/)","17.86 (min 17.19, max 19.19)" + "ippktcheck-4m (workloads/)","401.16 (min 390.44, max 420.95)" + "ippktcheck-4mw1 (workloads/)","400.93 (min 390.26, max 419.96)" + "ipres-4m (workloads/)","99.59 (min 97.78, max 103.02)" + "ipres-4mw1 (workloads/)","99.17 (min 97.40, max 102.39)" + "md5-4m (workloads/)","26.22 (min 25.27, max 28.03)" + "md5-4mw1 (workloads/)","26.08 (min 25.09, max 27.90)" + "rgbcmyk-4m (workloads/)","61.17 (min 58.82, max 65.83)" + "rgbcmyk-4mw1 (workloads/)","61.16 (min 58.81, max 65.77)" + "rotate-4ms1 (workloads/)","22.01 (min 21.25, max 23.46)" + "rotate-4ms1w1 (workloads/)","22.03 (min 21.29, max 23.42)" + "rotate-4ms64 (workloads/)","22.19 (min 21.35, max 23.67)" + "rotate-4ms64w1 (workloads/)","22.24 (min 21.51, max 23.63)" + "x264-4mq (workloads/)","0.54 (min 0.51, max 0.58)" + "x264-4mqw1 (workloads/)","0.54 (min 0.52, max 0.58)" + +| Boot-time Measurement ---------------------- +===================== Boot media: MMCSD -^^^^^^^^^^^^^^^^^ +----------------- .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am62pxx_sk-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","13.59 (min 13.10, max 18.72)" + "Linux boot time from SD with default rootfs (20 boot cycles)","17.86 (min 16.83, max 19.02)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. | -ALSA SoC Audio Driver ---------------------- +ALSA SoC Audio +============== #. Access type - RW\_INTERLEAVED #. Channels - 2 @@ -383,62 +395,63 @@ ALSA SoC Audio Driver .. csv-table:: Audio Capture :header: "Sampling Rate (Hz)","am62pxx_sk-fs: Throughput (bits/sec)","am62pxx_sk-fs: CPU Load (%)" - "11025","352798.33 (min 352797.00, max 352800.00)","0.10 (min 0.08, max 0.10)" - "16000","511998.83 (min 511997.00, max 512002.00)","0.23 (min 0.08, max 0.48)" - "22050","705594.83 (min 705589.00, max 705599.00)","0.13 (min 0.11, max 0.15)" - "24000","705597.17 (min 705595.00, max 705599.00)","0.14 (min 0.13, max 0.15)" - "32000","1023996.67 (min 1023994.00, max 1023998.00)","0.23 (min 0.08, max 0.87)" - "44100","1411195.33 (min 1411192.00, max 1411197.00)","0.21 (min 0.20, max 0.23)" - "48000","1535995.17 (min 1535991.00, max 1535998.00)","0.26 (min 0.09, max 0.92)" - "88200","2822388.00 (min 2822381.00, max 2822394.00)","0.38 (min 0.35, max 0.40)" - "96000","3071985.50 (min 3071973.00, max 3071993.00)","0.19 (min 0.17, max 0.21)" + "11025","352798.20 (min 352797.00, max 352799.00)","0.12 (min 0.10, max 0.13)" + "16000","511998.20 (min 511996.00, max 512000.00)","0.11 (min 0.10, max 0.11)" + "22050","705595.60 (min 705594.00, max 705597.00)","0.14 (min 0.13, max 0.15)" + "24000","705596.80 (min 705594.00, max 705599.00)","0.16 (min 0.14, max 0.18)" + "32000","1023995.60 (min 1023991.00, max 1023999.00)","0.38 (min 0.09, max 1.50)" + "44100","1411193.40 (min 1411187.00, max 1411199.00)","0.23 (min 0.20, max 0.24)" + "48000","1535993.00 (min 1535986.00, max 1535999.00)","0.21 (min 0.09, max 0.65)" + "88200","2822383.20 (min 2822371.00, max 2822395.00)","0.37 (min 0.35, max 0.40)" + "96000","3071966.00 (min 3071934.00, max 3071992.00)","0.27 (min 0.12, max 0.71)" .. csv-table:: Audio Playback :header: "Sampling Rate (Hz)","am62pxx_sk-fs: Throughput (bits/sec)","am62pxx_sk-fs: CPU Load (%)" - "11025","352946.33 (min 352945.00, max 352947.00)","0.09 (min 0.07, max 0.10)" - "16000","512213.17 (min 512210.00, max 512215.00)","0.08 (min 0.05, max 0.10)" - "22050","705849.17 (min 705824.00, max 705896.00)","0.11 (min 0.09, max 0.12)" - "24000","705891.33 (min 705888.00, max 705896.00)","0.13 (min 0.11, max 0.15)" - "32000","551612.00 (min 78800.00, max 1024427.00)","0.21 (min 0.05, max 0.84)" - "44100","1411628.17 (min 1411540.00, max 1411789.00)","0.18 (min 0.16, max 0.19)" - "48000","1536632.67 (min 1536629.00, max 1536635.00)","0.26 (min 0.12, max 0.51)" - "88200","2823553.00","0.34" - "96000","3073255.00","0.94" + "11025","352946.00 (min 352945.00, max 352947.00)","0.10 (min 0.09, max 0.11)" + "16000","512213.40 (min 512212.00, max 512215.00)","0.11 (min 0.10, max 0.12)" + "22050","705877.40 (min 705825.00, max 705892.00)","0.12" + "24000","705892.80 (min 705891.00, max 705895.00)","0.13 (min 0.12, max 0.14)" + "32000","835300.40 (min 78800.00, max 1024428.00)","0.09 (min 0.07, max 0.11)" + "44100","1411741.00 (min 1411558.00, max 1411790.00)","0.17 (min 0.15, max 0.18)" + "48000","1536638.50 (min 1536636.00, max 1536641.00)","0.10 (min 0.10, max 0.11)" + "88200","2823571.00 (min 2823566.00, max 2823576.00)","0.28 (min 0.24, max 0.31)" + "96000","3073263.00 (min 3073241.00, max 3073275.00)","0.14 (min 0.13, max 0.15)" | -Graphics SGX/RGX Driver ------------------------ +Graphics SGX/RGX +================ GFXBench -^^^^^^^^ +-------- + Run GFXBench and capture performance reported (Score and Display rate in fps). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: GFXBench Performance :header: "Benchmark","am62pxx_sk-fs: Score","am62pxx_sk-fs: Fps" - " GFXBench 3.x gl_manhattan_off","906.46 (min 902.42, max 910.12)","14.62 (min 14.56, max 14.68)" - " GFXBench 3.x gl_trex_off","1582.33 (min 1576.54, max 1589.72)","28.26 (min 28.15, max 28.39)" - " GFXBench 4.x gl_4_off","260.66 (min 259.90, max 261.32)","4.41 (min 4.40, max 4.42)" - " GFXBench 5.x gl_5_high_off","114.64 (min 113.73, max 115.05)","1.78 (min 1.77, max 1.79)" + " GFXBench 3.x gl_manhattan_off","863.66 (min 842.62, max 903.54)","13.93 (min 13.59, max 14.57)" + " GFXBench 3.x gl_trex_off","1591.39","28.42" + " GFXBench 5.x gl_5_high_off","108.08 (min 104.61, max 115.03)","1.68 (min 1.63, max 1.79)" Glmark2 -^^^^^^^ +------- Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: Glmark2 Performance :header: "Benchmark","am62pxx_sk-fs: Score" - "Glmark2-DRM","314.14 (min 307.00, max 354.00)" - "Glmark2-Wayland","730.75 (min 717.00, max 746.00)" + "Glmark2-DRM","276.50 (min 157.00, max 353.00)" + "Glmark2-Wayland","733.20 (min 718.00, max 782.00)" | Ethernet ------------------ -Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html +======== + +Ethernet performance benchmarks were measured using :command:`netperf` 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, @@ -448,11 +461,11 @@ CPU utilization was captured as the total percentage used across all cores on th while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. -In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth +In this scenario, :command:`netperf` options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: -:: +.. code-block:: console burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) burst_size = 500000000 / 8 / 1472 / 100 = 425 @@ -460,21 +473,21 @@ no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B data wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the :command:`netperf` test with no bandwidth limit (remove -b/-w options). -In order to start a netperf client on one device, the other device must have netserver running. -To start netserver: +In order to start a :command:`netperf` client on one device, the other device must have :command:`netserver` running. +To start :command:`netserver`: -:: +.. code-block:: console netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger :command:`netperf` clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test. -:: +.. code-block:: console #!/bin/bash for i in 1 @@ -486,87 +499,95 @@ additional timing measurements during the test. -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for +Running the following commands will trigger :command:`netperf` clients to measure UDP burst performance for 60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run :command:`netperf` client from DUT and start :command:`netserver` on tester. -:: +.. code-block:: console netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run :command:`netperf` client from tester and start :command:`netserver` on DUT. -:: +.. code-block:: console netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE +CPSW/CPSW2g/CPSW3g Ethernet +--------------------------- -CPSW/CPSW2g/CPSW3g Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -- CPSW2g: AM65x, J7200, J721e, J721S2, J784S4, J742S2 -- CPSW3g: AM64x, AM62x, AM62ax, AM62px +- CPSW3g: AM62px -.. rubric:: TCP Bidirectional Throughput - :name: CPSW2g-tcp-bidirectional-throughput +TCP Bidirectional Throughput +^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1778.39 (min 1503.06, max 1857.97)","63.69 (min 60.80, max 65.40)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1553.15 (min 1176.25, max 1811.29)","54.49 (min 39.05, max 70.68)" -.. rubric:: TCP Bidirectional Throughput Interrupt Pacing - :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing +TCP Bidirectional Throughput Interrupt Pacing +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1789.08 (min 1612.03, max 1873.08)","35.80 (min 27.35, max 38.76)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1849.66 (min 1825.39, max 1875.16)","38.86 (min 32.01, max 45.30)" -.. rubric:: UDP Throughput - :name: CPSW2g-udp-throughput-0-loss +UDP Throughput +^^^^^^^^^^^^^^ .. csv-table:: CPSW2g UDP Egress Throughput 0 loss :header: "Frame Size(bytes)","am62pxx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: Packets Per Second (kPPS)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","54.23 (min 53.64, max 55.42)","105.88 (min 105.00, max 108.00)","39.59 (min 39.25, max 40.20)" - "128","82.00","107.70 (min 106.76, max 110.02)","105.00 (min 104.00, max 107.00)","39.41 (min 39.06, max 39.88)" - "256","210.00","214.09 (min 211.33, max 217.17)","104.50 (min 103.00, max 106.00)","39.31 (min 39.08, max 39.69)" - "1024","978.00","836.06 (min 813.57, max 847.48)","101.88 (min 99.00, max 103.00)","39.65 (min 38.57, max 40.42)" - "1518","1472.00","838.69 (min 826.64, max 852.20)","69.13 (min 68.00, max 70.00)","37.17 (min 36.51, max 38.15)" + "64","","51.17 (min 41.42, max 63.85)","100.00 (min 81.00, max 125.00)","40.18 (min 37.30, max 43.08)" + "128","","101.30 (min 81.62, max 127.24)","99.00 (min 80.00, max 124.00)","39.93 (min 37.23, max 42.60)" + "256","","199.81 (min 162.14, max 250.72)","97.25 (min 79.00, max 122.00)","39.74 (min 36.74, max 42.59)" + "1024","","768.11 (min 634.77, max 935.43)","93.50 (min 77.00, max 114.00)","40.08 (min 36.82, max 45.20)" + "1518","","759.63 (min 622.71, max 917.85)","62.75 (min 51.00, max 76.00)","37.32 (min 34.57, max 41.16)" -| +.. csv-table:: CPSW2g UDP Ingress Throughput 0 loss + :header: "Frame Size(bytes)","am62pxx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: Packets Per Second (kPPS)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" -Linux OSPI Flash Driver ------------------------ + "128","","4.37 (min 3.89, max 5.02)","4.33 (min 4.00, max 5.00)","1.36 (min 0.83, max 2.25)" -.. rubric:: AM62PXX-SK - :name: am62pxx-sk-ospi +.. csv-table:: CPSW2g UDP Ingress Throughput possible loss + :header: "Frame Size(bytes)","am62pxx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: Packets Per Second (kPPS)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62pxx_sk-fs: Packet Loss %" -.. rubric:: UBIFS - :name: am62pxx-sk-ospi-ubifs + "128","","145.25 (min 124.88, max 181.77)","142.00 (min 122.00, max 178.00)","36.41 (min 34.88, max 39.17)","76.69 (min 71.36, max 80.26)" + +| + +OSPI +==== + +UBIFS +----- .. csv-table:: OSPI Flash Driver :header: "Buffer size (bytes)","am62pxx_sk-fs: Write UBIFS Throughput (Mbytes/sec)","am62pxx_sk-fs: Write UBIFS CPU Load (%)","am62pxx_sk-fs: Read UBIFS Throughput (Mbytes/sec)","am62pxx_sk-fs: Read UBIFS CPU Load (%)" - "102400","0.17 (min 0.12, max 0.28)","28.77 (min 24.69, max 33.84)","28.45 (min 28.34, max 28.64)","7.08 (min 3.57, max 12.50)" - "262144","0.14 (min 0.10, max 0.18)","29.70 (min 25.14, max 35.52)","28.38 (min 28.03, max 28.59)","8.42 (min 3.45, max 12.90)" - "524288","0.14 (min 0.10, max 0.18)","29.68 (min 26.42, max 33.97)","28.23 (min 27.97, max 28.56)","7.52 (min 3.45, max 12.50)" - "1048576","0.14 (min 0.10, max 0.18)","29.89 (min 26.72, max 33.36)","27.96 (min 27.72, max 28.27)","7.96 (min 6.67, max 9.68)" + "102400","0.17 (min 0.11, max 0.28)","28.43 (min 24.61, max 34.68)","27.63 (min 27.23, max 28.20)","9.70 (min 6.45, max 12.90)" + "262144","0.13 (min 0.10, max 0.18)","29.86 (min 25.97, max 34.50)","27.60 (min 27.18, max 28.27)","5.98 (min 3.45, max 6.67)" + "524288","0.13 (min 0.10, max 0.18)","29.15 (min 25.26, max 32.64)","27.37 (min 27.03, max 27.85)","6.63 (min 3.45, max 9.68)" + "1048576","0.13 (min 0.10, max 0.18)","30.02 (min 27.22, max 33.36)","27.16 (min 26.74, max 27.82)","9.50 (min 3.33, max 12.90)" -.. rubric:: RAW - :name: am62pxx-sk-ospi-raw +OSPI RAW +-------- .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62pxx_sk-fs: Raw Read Throughput (Mbytes/sec)" - "50","37.79 (min 37.31, max 37.88)" + "50","37.88" + +| + +EMMC +==== -EMMC Driver ------------ .. warning:: **IMPORTANT**: The performance numbers can be severely affected if the media is @@ -576,62 +597,51 @@ EMMC Driver re-mount in async mode. EMMC EXT4 FIO 1G -^^^^^^^^^^^^^^^^ +---------------- .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "1m","91.41 (min 90.40, max 92.50)","1.55 (min 1.48, max 1.61)","284.00 (min 276.00, max 289.00)","2.61 (min 2.33, max 2.79)" - "4m","95.41 (min 90.40, max 97.50)","1.06 (min 1.00, max 1.13)","230.14 (min 153.00, max 288.00)","1.72 (min 1.26, max 2.10)" - "4k","79.04 (min 78.90, max 79.30)","25.61 (min 25.54, max 25.65)","91.77 (min 89.20, max 93.10)","21.42 (min 20.63, max 21.94)" - "256k","91.33 (min 90.70, max 91.80)","1.97 (min 1.84, max 2.09)","288.43 (min 280.00, max 291.00)","4.00 (min 3.76, max 4.15)" - -EMMC RAW FIO 1G -^^^^^^^^^^^^^^^ - -.. csv-table:: EMMC RAW FIO 1G - :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - - "1m","90.53 (min 89.80, max 91.30)","1.37 (min 1.21, max 1.54)","293.29 (min 292.00, max 294.00)","2.64 (min 2.37, max 2.89)" - "4m","96.93 (min 96.50, max 97.50)","0.97 (min 0.86, max 1.03)","277.00 (min 182.00, max 294.00)","1.90 (min 1.48, max 2.09)" - "4k","76.76 (min 76.40, max 77.10)","18.94 (min 18.79, max 19.25)","92.90 (min 92.50, max 93.10)","20.21 (min 19.71, max 20.64)" - "256k","90.53 (min 89.70, max 91.40)","1.73 (min 1.53, max 1.95)","293.57 (min 293.00, max 294.00)","3.83 (min 3.73, max 3.91)" + "1m","91.50 (min 90.50, max 92.60)","1.68 (min 1.62, max 1.76)","265.60 (min 172.00, max 292.00)","2.75 (min 1.76, max 3.10)" + "4m","96.74 (min 96.20, max 97.20)","1.06 (min 0.97, max 1.13)","259.00 (min 172.00, max 289.00)","2.00 (min 1.35, max 2.34)" + "4k","75.68 (min 63.50, max 79.10)","24.68 (min 20.29, max 26.37)","91.88 (min 88.50, max 94.10)","21.96 (min 21.39, max 22.51)" + "256k","91.10 (min 90.60, max 91.70)","2.08 (min 1.91, max 2.26)","267.80 (min 173.00, max 294.00)","4.07 (min 2.71, max 4.75)" EMMC EXT4 -^^^^^^^^^ +--------- .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "102400","83.27 (min 52.29, max 89.46)","5.94 (min 4.97, max 8.11)","178.78 (min 178.64, max 178.91)","8.44 (min 6.58, max 10.13)" - "262144","80.06 (min 49.05, max 89.23)","5.71 (min 3.40, max 8.43)","179.72 (min 172.91, max 181.11)","9.36 (min 8.77, max 9.91)" - "524288","75.36 (min 49.32, max 89.16)","5.33 (min 3.06, max 8.14)","180.13 (min 176.40, max 181.89)","8.74 (min 8.26, max 9.87)" - "1048576","75.65 (min 47.53, max 88.93)","5.37 (min 2.72, max 8.45)","181.47 (min 180.95, max 181.87)","8.62 (min 7.86, max 9.44)" - "5242880","76.21 (min 49.34, max 89.86)","5.37 (min 3.02, max 7.98)","180.45 (min 174.59, max 181.97)","8.71 (min 8.30, max 9.52)" + "102400","91.67 (min 87.34, max 94.14)","3.71 (min 2.83, max 5.23)","177.06 (min 168.93, max 179.17)","6.08 (min 5.15, max 6.84)" + "262144","81.61 (min 48.84, max 94.77)","2.36 (min 1.28, max 3.85)","179.77 (min 175.01, max 181.42)","5.75 (min 4.74, max 6.47)" + "524288","78.37 (min 47.78, max 94.13)","2.30 (min 1.14, max 4.65)","181.00 (min 177.82, max 182.35)","5.77 (min 4.76, max 6.47)" + "1048576","77.61 (min 48.31, max 94.36)","2.28 (min 1.14, max 3.87)","180.95 (min 178.16, max 182.00)","5.19 (min 4.78, max 6.03)" + "5242880","75.53 (min 47.37, max 95.20)","2.43 (min 1.12, max 4.32)","181.68 (min 181.18, max 182.02)","5.64 (min 5.19, max 6.49)" EMMC VFAT -^^^^^^^^^ +--------- .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62pxx_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62pxx_sk-fs: Write VFAT CPU Load (%)","am62pxx_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62pxx_sk-fs: Read VFAT CPU Load (%)" - "102400","39.61 (min 11.80, max 52.28)","5.21 (min 3.66, max 6.47)","208.97 (min 208.62, max 209.38)","11.07 (min 9.23, max 12.81)" - "262144","44.50 (min 12.47, max 62.38)","6.00 (min 4.12, max 9.12)","287.77 (min 285.96, max 289.30)","15.04 (min 12.77, max 16.55)" - "524288","50.98 (min 12.69, max 73.49)","5.12 (min 3.75, max 7.69)","287.10 (min 285.81, max 288.26)","15.14 (min 14.58, max 16.11)" - "1048576","53.77 (min 12.83, max 75.68)","5.03 (min 3.49, max 6.69)","287.01 (min 285.60, max 288.35)","14.66 (min 13.29, max 15.75)" - "5242880","55.67 (min 13.01, max 81.71)","5.12 (min 3.81, max 6.94)","286.37 (min 280.42, max 288.38)","13.94 (min 12.77, max 15.75)" + "102400","38.37 (min 11.31, max 52.60)","5.05 (min 3.66, max 7.13)","200.84 (min 174.06, max 210.23)","11.14 (min 8.40, max 13.24)" + "262144","44.05 (min 11.91, max 62.82)","5.83 (min 4.00, max 8.71)","263.27 (min 176.35, max 288.62)","16.52 (min 9.70, max 22.92)" + "524288","50.43 (min 12.09, max 73.37)","5.16 (min 3.42, max 6.98)","263.58 (min 176.08, max 288.67)","14.38 (min 8.90, max 16.44)" + "1048576","52.85 (min 12.24, max 75.55)","5.12 (min 3.58, max 6.94)","262.95 (min 176.25, max 285.29)","14.56 (min 8.05, max 17.57)" + "5242880","55.67 (min 12.34, max 82.40)","5.28 (min 3.61, max 7.06)","262.72 (min 175.95, max 284.89)","14.16 (min 9.62, max 16.44)" -UBoot EMMC Driver ------------------ +UBoot EMMC +---------- .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62pxx_sk-fs: Write Throughput (Kbytes/sec)","am62pxx_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","96461.77 (min 93891.12, max 100515.34)","215037.67 (min 146285.71, max 282482.76)" - "4000000","97228.73 (min 95812.87, max 99447.65)","259335.04 (min 229950.88, max 296542.99)" + "2000000","98095.04 (min 95255.81, max 101135.80)","196449.26 (min 155298.58, max 268590.16)" + "4000000","97568.90 (min 95672.99, max 99598.78)","247676.85 (min 193893.49, max 302009.22)" MMCSD ------ +===== .. warning:: @@ -642,38 +652,27 @@ MMCSD re-mount in async mode. MMC EXT4 FIO 1G -^^^^^^^^^^^^^^^ +--------------- .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "1m","42.59 (min 41.40, max 43.40)","1.06 (min 1.01, max 1.16)","87.31 (min 87.20, max 87.40)","1.29 (min 1.15, max 1.39)" - "4m","42.17 (min 41.40, max 43.30)","0.71 (min 0.64, max 0.76)","87.24 (min 87.10, max 87.30)","0.92 (min 0.87, max 0.98)" - "4k","2.80 (min 2.78, max 2.83)","1.67 (min 1.60, max 1.75)","12.94 (min 12.80, max 13.20)","4.31 (min 4.01, max 4.52)" - "256k","38.44 (min 36.60, max 39.10)","1.25 (min 1.20, max 1.31)","83.47 (min 83.20, max 83.80)","1.52 (min 1.44, max 1.61)" - -MMC RAW FIO 1G -^^^^^^^^^^^^^^ - -.. csv-table:: MMC RAW FIO 1G - :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - - "1m","43.59 (min 42.40, max 45.10)","0.91 (min 0.86, max 1.01)","88.19 (min 88.10, max 88.30)","1.14 (min 1.09, max 1.22)" - "4m","43.20 (min 42.10, max 45.00)","0.70 (min 0.63, max 0.76)","88.14 (min 88.00, max 88.30)","0.90 (min 0.84, max 0.96)" - "4k","2.81 (min 2.80, max 2.83)","1.38 (min 1.33, max 1.43)","13.06 (min 13.00, max 13.10)","3.91 (min 3.82, max 4.06)" - "256k","38.00 (min 36.20, max 41.40)","1.02 (min 0.96, max 1.08)","84.33 (min 84.20, max 84.50)","1.44 (min 1.38, max 1.50)" + "1m","43.06 (min 42.00, max 43.70)","1.06 (min 0.97, max 1.13)","87.44 (min 87.10, max 88.20)","1.34 (min 1.23, max 1.42)" + "4m","42.16 (min 40.90, max 43.30)","0.69 (min 0.61, max 0.75)","86.26 (min 82.30, max 87.40)","0.94 (min 0.83, max 1.06)" + "4k","2.83 (min 2.79, max 2.88)","1.66 (min 1.51, max 1.76)","13.10 (min 12.90, max 13.50)","4.37 (min 3.98, max 4.93)" + "256k","38.92 (min 38.50, max 39.50)","1.30 (min 1.13, max 1.44)","83.68 (min 83.30, max 84.40)","1.59 (min 1.47, max 1.72)" MMC EXT4 -^^^^^^^^ +-------- .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - "102400","10.75 (min 10.41, max 11.18)","0.83 (min 0.67, max 1.20)","11.19 (min 10.89, max 11.82)","0.75 (min 0.68, max 0.90)" - "262144","10.43 (min 10.25, max 10.62)","0.90 (min 0.68, max 1.46)","11.14 (min 10.99, max 11.22)","0.73 (min 0.63, max 0.78)" - "524288","10.46 (min 10.29, max 10.86)","0.81 (min 0.65, max 1.20)","11.10 (min 10.85, max 11.50)","0.68 (min 0.61, max 0.88)" - "1048576","10.49 (min 10.27, max 10.76)","0.88 (min 0.68, max 1.18)","11.06 (min 10.85, max 11.50)","0.66 (min 0.58, max 0.74)" - "5242880","10.64 (min 10.27, max 11.12)","0.83 (min 0.64, max 1.20)","11.59 (min 10.97, max 12.03)","0.66 (min 0.58, max 0.73)" + "102400","10.84 (min 10.58, max 11.28)","0.48 (min 0.36, max 0.69)","11.30 (min 10.87, max 11.81)","0.49 (min 0.45, max 0.61)" + "262144","10.82 (min 10.65, max 11.18)","0.41 (min 0.29, max 0.73)","11.09 (min 10.97, max 11.24)","0.57 (min 0.50, max 0.68)" + "524288","10.85 (min 10.31, max 11.10)","0.38 (min 0.26, max 0.55)","11.11 (min 10.88, max 11.52)","0.45 (min 0.39, max 0.53)" + "1048576","10.83 (min 10.11, max 11.33)","0.43 (min 0.29, max 0.62)","11.67 (min 11.27, max 12.03)","0.49 (min 0.38, max 0.63)" + "5242880","10.99 (min 10.47, max 11.35)","0.39 (min 0.26, max 0.59)","11.83 (min 11.06, max 12.03)","0.43 (min 0.40, max 0.49)" The performance numbers were captured using the following: @@ -689,9 +688,9 @@ UBOOT MMCSD FAT .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62pxx_sk-fs: Write Throughput (Kbytes/sec)","am62pxx_sk-fs: Read Throughput (Kbytes/sec)" - "400000","36689.07 (min 28054.79, max 43574.47)","82875.34 (min 81920.00, max 83591.84)" - "800000","40145.70 (min 33032.26, max 47627.91)","87148.94" - "1000000","48396.32 (min 42555.84, max 51360.50)","89670.60 (min 89530.05, max 90021.98)" + "400000","36090.73 (min 33573.77, max 40960.00)","82588.74 (min 81920.00, max 83591.84)" + "800000","41443.90 (min 34276.15, max 43807.49)","87148.94" + "1000000","46962.65 (min 42226.80, max 49498.49)","89530.05" The performance numbers were captured using the following: @@ -699,93 +698,93 @@ The performance numbers were captured using the following: | -USB Driver ----------- +USB +=== USB Device Controller -^^^^^^^^^^^^^^^^^^^^^ +--------------------- .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT :header: "Number of Blocks","am62pxx_sk-fs: Throughput (MB/sec)" - "150","30.13 (min 26.60, max 43.90)" + "150","35.96 (min 27.00, max 38.60)" .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT :header: "Number of Blocks","am62pxx_sk-fs: Throughput (MB/sec)" - "150","25.53 (min 21.90, max 37.10)" + "150","30.72 (min 24.80, max 33.40)" | -CRYPTO Driver -------------- +CRYPTO +====== OpenSSL Performance -^^^^^^^^^^^^^^^^^^^ +------------------- .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am62pxx_sk-fs: throughput (KBytes/Sec)" - "aes-128-cbc","1024","23874.22 (min 22797.65, max 24305.66)" - "aes-128-cbc","16","433.40 (min 414.08, max 439.83)" - "aes-128-cbc","16384","86033.94 (min 84885.50, max 86523.90)" - "aes-128-cbc","256","7133.66 (min 6779.65, max 7281.83)" - "aes-128-cbc","64","1854.41 (min 1809.17, max 1895.91)" - "aes-128-cbc","8192","72677.86 (min 71183.02, max 73187.33)" - "aes-128-ecb","1024","24521.24 (min 23560.19, max 24830.98)" - "aes-128-ecb","16","444.84 (min 423.89, max 450.61)" - "aes-128-ecb","16384","88832.49 (min 87801.86, max 89511.25)" - "aes-128-ecb","256","7297.62 (min 6986.33, max 7385.43)" - "aes-128-ecb","64","1921.89 (min 1846.42, max 1959.10)" - "aes-128-ecb","8192","74998.15 (min 72996.18, max 75606.70)" - "aes-192-cbc","1024","23428.24 (min 22448.13, max 23803.22)" - "aes-192-cbc","16","433.98 (min 413.37, max 440.67)" - "aes-192-cbc","16384","77998.76 (min 76961.11, max 78419.29)" - "aes-192-cbc","256","7084.75 (min 6741.16, max 7235.16)" - "aes-192-cbc","64","1858.29 (min 1806.49, max 1907.71)" - "aes-192-cbc","8192","66915.77 (min 65544.19, max 67439.27)" - "aes-192-ecb","1024","24060.10 (min 23176.19, max 24450.05)" - "aes-192-ecb","16","444.52 (min 424.39, max 449.94)" - "aes-192-ecb","16384","80076.41 (min 78779.73, max 80767.66)" - "aes-192-ecb","256","7282.65 (min 6920.45, max 7394.73)" - "aes-192-ecb","64","1914.58 (min 1845.12, max 1949.12)" - "aes-192-ecb","8192","68555.34 (min 67365.55, max 69383.51)" - "aes-256-cbc","1024","22864.60 (min 21938.52, max 23255.72)" - "aes-256-cbc","16","435.49 (min 413.74, max 442.78)" - "aes-256-cbc","16384","70860.02 (min 69817.69, max 71636.31)" - "aes-256-cbc","256","7052.84 (min 6704.38, max 7207.94)" - "aes-256-cbc","64","1867.00 (min 1805.76, max 1916.16)" - "aes-256-cbc","8192","61503.98 (min 60598.95, max 62406.66)" - "aes-256-ecb","1024","23521.77 (min 22587.39, max 23857.15)" - "aes-256-ecb","16","444.28 (min 425.70, max 450.73)" - "aes-256-ecb","16384","72667.72 (min 72176.98, max 73558.70)" - "aes-256-ecb","256","7234.73 (min 6907.56, max 7374.17)" - "aes-256-ecb","64","1913.51 (min 1846.23, max 1957.03)" - "aes-256-ecb","8192","63232.10 (min 62390.27, max 63744.68)" - "sha256","1024","37936.08 (min 37196.12, max 38420.14)" - "sha256","16","632.25 (min 614.09, max 639.77)" - "sha256","16384","299548.67 (min 296605.01, max 302246.57)" - "sha256","256","9906.47 (min 9659.73, max 10023.59)" - "sha256","64","2505.28 (min 2431.06, max 2528.36)" - "sha256","8192","203446.37 (min 199546.20, max 205512.70)" - "sha512","1024","26078.79 (min 25618.77, max 26265.94)" - "sha512","16","612.27 (min 590.90, max 623.37)" - "sha512","16384","68343.91 (min 67993.60, max 68517.89)" - "sha512","256","8707.43 (min 8436.31, max 8843.78)" - "sha512","64","2455.43 (min 2363.48, max 2505.77)" - "sha512","8192","61529.72 (min 61104.13, max 61680.30)" + "aes-128-cbc","1024","22662.57 (min 21330.94, max 24852.48)" + "aes-128-cbc","16","419.93 (min 405.77, max 451.41)" + "aes-128-cbc","16384","85168.13 (min 84497.75, max 86633.13)" + "aes-128-cbc","256","6960.34 (min 6771.88, max 7385.94)" + "aes-128-cbc","64","1841.70 (min 1769.47, max 1961.96)" + "aes-128-cbc","8192","71877.29 (min 71005.53, max 73714.35)" + "aes-128-ecb","1024","23194.71 (min 21783.55, max 25396.91)" + "aes-128-ecb","16","429.13 (min 415.93, max 459.21)" + "aes-128-ecb","16384","88210.09 (min 87332.18, max 89571.33)" + "aes-128-ecb","256","7081.71 (min 6928.81, max 7419.65)" + "aes-128-ecb","64","1889.19 (min 1835.82, max 2021.99)" + "aes-128-ecb","8192","74016.09 (min 73064.45, max 76046.34)" + "aes-192-cbc","1024","22294.78 (min 20896.77, max 24468.82)" + "aes-192-cbc","16","421.20 (min 406.44, max 453.28)" + "aes-192-cbc","16384","77471.74 (min 76819.11, max 78697.81)" + "aes-192-cbc","256","6940.52 (min 6709.25, max 7392.43)" + "aes-192-cbc","64","1848.60 (min 1800.21, max 1973.61)" + "aes-192-cbc","8192","66239.83 (min 65129.13, max 68182.02)" + "aes-192-ecb","1024","22754.90 (min 21388.97, max 24982.19)" + "aes-192-ecb","16","429.45 (min 415.78, max 460.86)" + "aes-192-ecb","16384","79107.41 (min 78129.83, max 80013.99)" + "aes-192-ecb","256","7031.79 (min 6869.93, max 7419.90)" + "aes-192-ecb","64","1887.06 (min 1827.56, max 2022.78)" + "aes-192-ecb","8192","67811.33 (min 67212.63, max 68577.96)" + "aes-256-cbc","1024","21804.80 (min 20580.35, max 23639.04)" + "aes-256-cbc","16","421.40 (min 407.79, max 452.66)" + "aes-256-cbc","16384","70429.35 (min 69533.70, max 71286.78)" + "aes-256-cbc","256","6910.42 (min 6666.84, max 7359.06)" + "aes-256-cbc","64","1847.16 (min 1798.85, max 1968.02)" + "aes-256-cbc","8192","61151.91 (min 60593.49, max 61991.59)" + "aes-256-ecb","1024","22310.49 (min 21001.90, max 24208.38)" + "aes-256-ecb","16","429.48 (min 414.78, max 460.38)" + "aes-256-ecb","16384","72333.99 (min 71374.17, max 73433.09)" + "aes-256-ecb","256","7017.88 (min 6817.88, max 7410.43)" + "aes-256-ecb","64","1884.53 (min 1836.84, max 2008.73)" + "aes-256-ecb","8192","62623.74 (min 61491.88, max 63670.95)" + "sha256","1024","29610.41 (min 28761.43, max 31727.96)" + "sha256","16","483.89 (min 461.47, max 527.91)" + "sha256","16384","258501.29 (min 250724.35, max 275256.66)" + "sha256","256","7645.93 (min 7375.87, max 8255.57)" + "sha256","64","1921.05 (min 1845.40, max 2081.13)" + "sha256","8192","169372.33 (min 164525.40, max 180974.93)" + "sha512","1024","24823.30 (min 23840.43, max 26991.96)" + "sha512","16","475.94 (min 460.72, max 517.66)" + "sha512","16384","101677.74 (min 98396.84, max 110455.47)" + "sha512","256","7190.93 (min 6925.57, max 7765.50)" + "sha512","64","1903.05 (min 1842.90, max 2071.62)" + "sha512","8192","84061.53 (min 81250.99, max 91379.03)" .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am62pxx_sk-fs: CPU Load" - "aes-128-cbc","32.71 (min 31.00, max 33.00)" - "aes-128-ecb","33.86 (min 32.00, max 35.00)" - "aes-192-cbc","32.71 (min 31.00, max 33.00)" - "aes-192-ecb","33.29 (min 32.00, max 34.00)" - "aes-256-cbc","31.57 (min 30.00, max 32.00)" - "aes-256-ecb","32.43 (min 31.00, max 33.00)" - "sha256","95.57 (min 94.00, max 96.00)" - "sha512","95.86 (min 95.00, max 96.00)" + "aes-128-cbc","34.25 (min 34.00, max 35.00)" + "aes-128-ecb","35.50 (min 35.00, max 36.00)" + "aes-192-cbc","34.25 (min 33.00, max 35.00)" + "aes-192-ecb","34.75 (min 34.00, max 35.00)" + "aes-256-cbc","33.50 (min 33.00, max 34.00)" + "aes-256-ecb","34.50 (min 34.00, max 35.00)" + "sha256","96.00" + "sha512","96.00" Listed for each algorithm are the code snippets used to run each benchmark test. @@ -794,81 +793,4 @@ Listed for each algorithm are the code snippets used to run each time -v openssl speed -elapsed -evp aes-128-cbc -Low Power Performance ---------------------- - -Power Performance -^^^^^^^^^^^^^^^^^ - -.. csv-table:: I/O Only + DDR Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core","0.85","0.00" - "vddr_core","0.85","0.00" - "soc_dvdd_3v3","3.30","2.07" - "soc_dvdd_1v8","1.80","8.50" - "vdda_1v8","1.80","0.13" - "vdd_lpddr4","1.10","4.58" - "Total"," ","15.28" - -.. csv-table:: Deep Sleep Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core","0.85","18.32" - "vddr_core","0.85","1.53" - "soc_dvdd_3v3","3.30","8.54" - "soc_dvdd_1v8","1.80","3.01" - "vdda_1v8","1.80","68.06" - "vdd_lpddr4","1.10","3.47" - "Total"," ","102.93" - -.. csv-table:: MCU Only Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core","0.85","219.30" - "vddr_core","0.85","3.32" - "soc_dvdd_3v3","3.30","7.10" - "soc_dvdd_1v8","1.80","2.93" - "vdda_1v8","1.80","76.55" - "vdd_lpddr4","1.10","4.37" - "Total"," ","313.56" - -Partial I/O Data -- All voltage rails were measured to be near 0V - -.. note:: - - The power consumption on the vdda_1v8 rail is not indicitive of the SoC's power consumption due to an oscillator on the rail - that has significant current consumption. - -.. note:: - - The measurements shown are from an AM62Px SK rev E1-1. Results may vary based off of the board revision being used. - -Further optimizations are possible for these low power modes. Please refer to the AM62x Power Consumption App Note (https://www.ti.com/lit/pdf/spradg1) - -Resume Latency Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^ - -.. csv-table:: LPM Resume Latency Performance - :header: "Low Power Mode","Total Resume Latency (ms)" - - "I/O Only + DDR", "709.82" - "Deep Sleep", "136.52" - "MCU Only", "91.68" - -The performance numbers are measured without the Linux printk logs. To remove the -Linux printk logs, run the following commands in the terminal: - -.. code:: console - - # Detach kernel serial console - consoles=$(find /sys/bus/platform/devices/*.serial/ -name console) - for console in ${consoles}; do - echo -n N > ${console} - done - - -.. note:: - - The measurements shown are from using the default SDK with no extra optimizations. +| diff --git a/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst index 3522b2e6e..60b522d39 100644 --- a/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst @@ -1,9 +1,10 @@ -======================================= - RT-linux 11.02.08.02 Performance Guide -======================================= +################################### +RT-linux 12.00.00 Performance Guide +################################### -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -16,8 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-rt-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -25,60 +27,13 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-rt-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ | +***************** System Benchmarks ------------------ - -Stress-ng and Cyclic Test - -.. _RT-linux-performance: - -stress-ng (next-generation) will stress test a embedded platform in various selectable ways. -It was designed to exercise various physical subsystems as well as the various -operating system kernel interfaces. stress-ng can also measure test throughput rates; -this can be useful to observe performance changes across different operating system or types of hardware. - -Cyclictest is most commonly used for benchmarking RT systems. -It is one of the most frequently used tools for evaluating the relative performance of real-time systems. -Some performance tests which use Cyclictest are System benchmarking, Latency debugging with tracing and -approximating application performance. - -Test commands used for running stress-ng and cyclictest together - -.. code:: console - - stress-ng --cpu-method=all -c 4 & - cyclictest -m -Sp80 -D6h -h400 -i200 -M -q +***************** -The following summarizes the latencies observed using the yocto based -default SDK image - -.. note:: - - Using the OP-TEE TRNG driver can impact this benchmark's performance due to - frequent context switching between Normal World (Linux) and Secure World (OP-TEE), - that occurs when the kernel's hardware random number generator interface - (hwrng) requests entropy from the secure TRNG to replenish the Linux entropy - pool. - - The Linux TRNG driver can mitigate these latency spikes. This involves - enabling the Pseudo RNG driver in OP-TEE as documented in the Foundational - Components section: :ref:`building-optee-with-prng`, and enabling the RNG - node in the Linux kernel device tree. This way the HW TRNG is accessed from - the kernel itself. - -.. csv-table:: - :header: "Latencies","CPU 0","CPU 1","CPU 2","CPU 3" - - "Minimum (usec)","5","5","5","5" - "Average (usec)","6","6","6","6" - "Maximum (usec)","33","36","34","40" - -.. image:: img/rt-cpu-method-all-latency-histogram.png +| diff --git a/source/devices/AM62X/linux/Linux_Performance_Guide.rst b/source/devices/AM62X/linux/Linux_Performance_Guide.rst index be0422e78..ab2394922 100644 --- a/source/devices/AM62X/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62X/linux/Linux_Performance_Guide.rst @@ -1,10 +1,11 @@ -==================================== - Linux 11.02.08.02 Performance Guide -==================================== +################################# + Linux 12.00.00 Performance Guide +################################# -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -21,8 +22,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -30,16 +32,20 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ +| + +***************** System Benchmarks ------------------ +***************** + +| + LMBench -^^^^^^^ +======= + LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at @@ -66,124 +72,125 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "af_unix_sock_stream_latency (microsec)","33.40 (min 32.86, max 34.65)","29.79 (min 28.64, max 30.84)","33.61 (min 33.07, max 34.90)" - "af_unix_socket_stream_bandwidth (mb\s)","532.62 (min 513.45, max 555.88)","602.54 (min 576.79, max 715.30)","728.89 (min 632.89, max 809.03)" - "bw_file_rd-io-1mb (mb/s)","844.54 (min 822.23, max 886.68)","985.59 (min 935.75, max 1129.31)","970.83 (min 868.06, max 1075.08)" - "bw_file_rd-o2c-1mb (mb/s)","460.27 (min 433.28, max 488.60)","520.63 (min 488.28, max 589.88)","545.10 (min 484.50, max 608.46)" - "bw_mem-bcopy-16mb (mb/s)","666.43 (min 656.63, max 698.69)","779.46 (min 755.22, max 916.43)","804.83 (min 710.04, max 926.57)" - "bw_mem-bcopy-1mb (mb/s)","700.75 (min 675.45, max 741.70)","842.36 (min 785.55, max 1038.96)","888.53 (min 745.85, max 1009.08)" - "bw_mem-bcopy-2mb (mb/s)","657.09 (min 645.27, max 687.52)","789.75 (min 751.88, max 965.41)","796.56 (min 719.55, max 885.48)" - "bw_mem-bcopy-4mb (mb/s)","689.78 (min 669.68, max 728.73)","816.64 (min 783.70, max 999.25)","813.86 (min 710.98, max 940.96)" - "bw_mem-bcopy-8mb (mb/s)","699.67 (min 684.46, max 732.80)","822.42 (min 799.44, max 920.70)","807.96 (min 708.84, max 914.81)" - "bw_mem-bzero-16mb (mb/s)","1706.95 (min 1678.03, max 1776.99)","1798.84 (min 1743.30, max 2125.68)","1932.00 (min 1724.88, max 2128.79)" - "bw_mem-bzero-1mb (mb/s)","1194.75 (min 675.45, max 1771.79)","1310.82 (min 785.55, max 2122.77)","1407.48 (min 745.85, max 2129.17)" - "bw_mem-bzero-2mb (mb/s)","1172.66 (min 645.27, max 1768.35)","1285.64 (min 751.88, max 2123.14)","1361.57 (min 719.55, max 2130.30)" - "bw_mem-bzero-4mb (mb/s)","1189.77 (min 669.68, max 1766.78)","1299.62 (min 783.70, max 2125.78)","1370.93 (min 710.98, max 2129.93)" - "bw_mem-bzero-8mb (mb/s)","1202.52 (min 684.46, max 1767.37)","1310.44 (min 799.44, max 2125.12)","1372.21 (min 708.84, max 2128.79)" - "bw_mem-cp-16mb (mb/s)","381.91 (min 356.51, max 400.99)","466.22 (min 438.89, max 574.16)","470.46 (min 411.22, max 525.11)" - "bw_mem-cp-1mb (mb/s)","1107.67 (min 369.41, max 1907.67)","1193.34 (min 436.81, max 2264.58)","1276.71 (min 394.48, max 2281.80)" - "bw_mem-cp-2mb (mb/s)","1071.32 (min 360.82, max 1842.30)","1160.88 (min 427.53, max 2201.03)","1231.90 (min 381.90, max 2205.48)" - "bw_mem-cp-4mb (mb/s)","1072.77 (min 398.29, max 1814.61)","1162.44 (min 467.24, max 2159.05)","1218.76 (min 398.45, max 2171.16)" - "bw_mem-cp-8mb (mb/s)","1071.48 (min 408.66, max 1804.85)","1164.03 (min 482.77, max 2152.85)","1203.39 (min 386.32, max 2174.80)" - "bw_mem-fcp-16mb (mb/s)","669.23 (min 654.21, max 689.36)","757.23 (min 728.96, max 882.76)","867.40 (min 755.89, max 1002.82)" - "bw_mem-fcp-1mb (mb/s)","1231.51 (min 756.43, max 1771.79)","1338.53 (min 855.29, max 2122.77)","1426.17 (min 823.32, max 2129.17)" - "bw_mem-fcp-2mb (mb/s)","1211.46 (min 718.13, max 1768.35)","1330.64 (min 832.52, max 2123.14)","1411.18 (min 782.78, max 2130.30)" - "bw_mem-fcp-4mb (mb/s)","1222.02 (min 729.13, max 1766.78)","1347.90 (min 869.28, max 2125.78)","1411.80 (min 780.03, max 2129.93)" - "bw_mem-fcp-8mb (mb/s)","1236.32 (min 736.85, max 1767.37)","1366.49 (min 887.71, max 2125.12)","1415.27 (min 788.18, max 2128.79)" - "bw_mem-frd-16mb (mb/s)","1022.55 (min 981.35, max 1076.35)","1188.74 (min 1132.50, max 1428.19)","1201.69 (min 1025.58, max 1350.67)" - "bw_mem-frd-1mb (mb/s)","958.09 (min 756.43, max 1193.93)","1114.58 (min 855.29, max 1598.58)","1143.78 (min 823.32, max 1534.53)" - "bw_mem-frd-2mb (mb/s)","871.89 (min 718.13, max 1081.28)","1010.04 (min 832.52, max 1378.12)","1048.22 (min 782.78, max 1358.00)" - "bw_mem-frd-4mb (mb/s)","886.22 (min 729.13, max 1068.09)","1041.72 (min 869.28, max 1402.28)","1046.94 (min 780.03, max 1353.41)" - "bw_mem-frd-8mb (mb/s)","895.53 (min 736.85, max 1082.98)","1056.14 (min 887.71, max 1422.22)","1050.26 (min 788.18, max 1356.62)" - "bw_mem-fwr-16mb (mb/s)","1713.36 (min 1690.44, max 1779.76)","1809.72 (min 1752.08, max 2139.90)","1946.08 (min 1751.89, max 2145.35)" - "bw_mem-fwr-1mb (mb/s)","1487.39 (min 1112.70, max 1907.67)","1624.29 (min 1263.54, max 2264.58)","1725.34 (min 1197.40, max 2281.80)" - "bw_mem-fwr-2mb (mb/s)","1388.30 (min 974.34, max 1842.30)","1502.10 (min 1075.85, max 2201.03)","1601.94 (min 1076.57, max 2205.48)" - "bw_mem-fwr-4mb (mb/s)","1378.25 (min 989.12, max 1814.61)","1502.80 (min 1074.98, max 2159.05)","1584.78 (min 1080.94, max 2171.16)" - "bw_mem-fwr-8mb (mb/s)","1373.81 (min 975.73, max 1804.85)","1498.37 (min 1119.66, max 2152.85)","1582.44 (min 1073.39, max 2174.80)" - "bw_mem-rd-16mb (mb/s)","1037.69 (min 1016.97, max 1083.72)","1187.07 (min 1098.90, max 1445.48)","1215.25 (min 1060.59, max 1351.12)" - "bw_mem-rd-1mb (mb/s)","867.03 (min 533.14, max 1234.35)","1028.36 (min 648.40, max 1634.88)","1048.08 (min 572.63, max 1595.18)" - "bw_mem-rd-2mb (mb/s)","767.51 (min 473.20, max 1085.19)","899.44 (min 569.48, max 1419.45)","923.31 (min 506.14, max 1376.70)" - "bw_mem-rd-4mb (mb/s)","798.39 (min 530.01, max 1083.13)","946.09 (min 674.76, max 1435.49)","922.27 (min 522.94, max 1364.26)" - "bw_mem-rd-8mb (mb/s)","831.83 (min 586.17, max 1084.60)","989.49 (min 743.43, max 1446.92)","937.43 (min 563.94, max 1356.62)" - "bw_mem-rdwr-16mb (mb/s)","670.46 (min 629.00, max 718.55)","836.08 (min 773.47, max 1046.44)","662.35 (min 584.13, max 754.68)" - "bw_mem-rdwr-1mb (mb/s)","463.26 (min 369.41, max 567.46)","576.99 (min 436.81, max 806.19)","571.73 (min 394.48, max 769.47)" - "bw_mem-rdwr-2mb (mb/s)","429.34 (min 360.82, max 514.07)","532.83 (min 427.53, max 740.60)","529.48 (min 381.90, max 690.25)" - "bw_mem-rdwr-4mb (mb/s)","480.32 (min 398.29, max 593.74)","588.73 (min 467.24, max 879.99)","548.02 (min 398.45, max 723.85)" - "bw_mem-rdwr-8mb (mb/s)","525.75 (min 408.66, max 655.68)","644.03 (min 482.77, max 956.71)","545.57 (min 386.32, max 737.87)" - "bw_mem-wr-16mb (mb/s)","667.31 (min 625.73, max 693.30)","844.62 (min 778.66, max 1039.37)","675.18 (min 587.59, max 771.12)" - "bw_mem-wr-1mb (mb/s)","545.02 (min 524.48, max 572.66)","690.54 (min 638.47, max 819.94)","676.66 (min 572.63, max 781.40)" - "bw_mem-wr-2mb (mb/s)","490.33 (min 468.06, max 525.97)","612.48 (min 567.54, max 740.60)","603.33 (min 506.14, max 707.21)" - "bw_mem-wr-4mb (mb/s)","557.39 (min 494.56, max 593.74)","697.74 (min 605.05, max 879.99)","626.41 (min 522.94, max 725.16)" - "bw_mem-wr-8mb (mb/s)","630.78 (min 586.17, max 671.76)","785.24 (min 709.66, max 1007.68)","648.27 (min 563.94, max 754.36)" - "bw_mmap_rd-mo-1mb (mb/s)","1163.27 (min 1139.38, max 1219.09)","1324.21 (min 1260.35, max 1606.86)","1377.07 (min 1225.06, max 1541.36)" - "bw_mmap_rd-o2c-1mb (mb/s)","455.11 (min 438.28, max 477.55)","500.79 (min 427.78, max 590.67)","519.99 (min 465.91, max 600.69)" - "bw_pipe (mb/s)","443.15 (min 435.10, max 453.94)","515.70 (min 492.51, max 600.38)","498.69 (min 451.80, max 545.64)" - "bw_unix (mb/s)","532.62 (min 513.45, max 555.88)","602.54 (min 576.79, max 715.30)","728.89 (min 632.89, max 809.03)" - "lat_connect (us)","66.67 (min 66.20, max 67.85)","59.11 (min 58.18, max 59.76)","65.67 (min 63.87, max 69.00)" - "lat_ctx-2-128k (us)","8.89 (min 8.32, max 9.14)","7.99 (min 7.63, max 8.34)","8.91 (min 8.50, max 9.59)" - "lat_ctx-2-256k (us)","13.57 (min 9.14, max 16.40)","9.88 (min 6.84, max 10.92)","12.44 (min 8.10, max 15.04)" - "lat_ctx-4-128k (us)","9.13 (min 8.33, max 9.61)","7.66 (min 7.17, max 8.17)","8.56 (min 8.09, max 8.88)" - "lat_ctx-4-256k (us)","10.98 (min 7.76, max 13.14)","9.48 (min 7.36, max 12.90)","8.46 (min 0.00, max 12.65)" - "lat_fs-0k (num_files)","198.86 (min 177.00, max 227.00)","213.63 (min 195.00, max 235.00)","198.88 (min 192.00, max 210.00)" - "lat_fs-10k (num_files)","92.57 (min 86.00, max 103.00)","102.88 (min 92.00, max 128.00)","93.13 (min 82.00, max 104.00)" - "lat_fs-1k (num_files)","124.29 (min 109.00, max 146.00)","145.63 (min 135.00, max 160.00)","134.75 (min 115.00, max 147.00)" - "lat_fs-4k (num_files)","123.14 (min 107.00, max 130.00)","145.25 (min 128.00, max 159.00)","133.63 (min 123.00, max 158.00)" - "lat_mem_rd-stride128-sz1000k (ns)","56.96 (min 54.29, max 58.01)","50.45 (min 42.65, max 52.51)","48.81 (min 43.55, max 54.17)" - "lat_mem_rd-stride128-sz125k (ns)","6.22 (min 6.20, max 6.26)","5.57 (min 5.54, max 5.58)","6.22 (min 6.20, max 6.26)" - "lat_mem_rd-stride128-sz250k (ns)","6.54 (min 6.54, max 6.56)","5.84 (min 5.83, max 5.85)","6.54 (min 6.53, max 6.54)" - "lat_mem_rd-stride128-sz31k (ns)","4.20 (min 2.42, max 4.70)","3.85 (min 3.65, max 4.19)","3.40 (min 2.41, max 4.70)" - "lat_mem_rd-stride128-sz50 (ns)","2.40 (min 2.40, max 2.41)","2.15","2.40 (min 2.40, max 2.41)" - "lat_mem_rd-stride128-sz500k (ns)","17.11 (min 11.46, max 21.66)","17.27 (min 13.44, max 20.26)","15.35 (min 10.09, max 21.82)" - "lat_mem_rd-stride128-sz62k (ns)","5.77 (min 5.05, max 5.89)","5.24 (min 5.22, max 5.27)","5.87 (min 5.85, max 5.90)" - "lat_mmap-1m (us)","66.14 (min 59.00, max 73.00)","55.13 (min 53.00, max 64.00)","64.75 (min 57.00, max 71.00)" - "lat_ops-double-add (ns)","3.21","2.86 (min 2.86, max 2.87)","3.21" - "lat_ops-double-div (ns)","17.63 (min 17.62, max 17.64)","15.74 (min 15.74, max 15.75)","17.63 (min 17.63, max 17.65)" - "lat_ops-double-mul (ns)","3.21","2.86","3.21" - "lat_ops-float-add (ns)","3.21","2.86 (min 2.86, max 2.87)","3.21" - "lat_ops-float-div (ns)","10.43 (min 10.42, max 10.43)","9.30","10.42 (min 10.42, max 10.43)" - "lat_ops-float-mul (ns)","3.21 (min 3.20, max 3.21)","2.86","3.21" - "lat_ops-int-add (ns)","0.80","0.72","0.80" - "lat_ops-int-bit (ns)","0.53","0.48","0.53 (min 0.53, max 0.54)" - "lat_ops-int-div (ns)","4.81 (min 4.81, max 4.82)","4.29 (min 4.29, max 4.30)","4.81" - "lat_ops-int-mod (ns)","5.08","4.53 (min 4.53, max 4.54)","5.08 (min 5.07, max 5.08)" - "lat_ops-int-mul (ns)","3.45 (min 3.45, max 3.46)","3.08 (min 3.07, max 3.10)","3.45 (min 3.44, max 3.46)" - "lat_ops-int64-add (ns)","0.80","0.72","0.80" - "lat_ops-int64-bit (ns)","0.53","0.48","0.53 (min 0.53, max 0.54)" - "lat_ops-int64-div (ns)","7.62 (min 7.61, max 7.62)","6.80","7.62 (min 7.61, max 7.62)" - "lat_ops-int64-mod (ns)","5.88","5.25","5.88 (min 5.87, max 5.89)" - "lat_ops-int64-mul (ns)","3.99 (min 3.98, max 4.04)","3.56 (min 3.55, max 3.61)","3.98" - "lat_pagefault (us)","0.80 (min 0.76, max 0.82)","0.71 (min 0.60, max 0.73)","1.26 (min 0.65, max 1.74)" - "lat_pipe (us)","28.82 (min 28.37, max 29.43)","25.77 (min 25.50, max 26.22)","28.93 (min 28.25, max 29.53)" - "lat_proc-exec (us)","1108.66 (min 1066.20, max 1142.00)","979.70 (min 844.00, max 1031.50)","1006.99 (min 886.29, max 1144.60)" - "lat_proc-fork (us)","940.39 (min 880.50, max 1009.50)","831.78 (min 684.14, max 873.14)","857.04 (min 765.86, max 954.67)" + "af_unix_sock_stream_latency (microsec)","29.86 (min 25.84, max 32.02)","25.62 (min 23.44, max 27.72)","27.27 (min 26.72, max 27.91)" + "af_unix_socket_stream_bandwidth (mb\s)","550.16 (min 537.95, max 572.90)","656.18 (min 589.74, max 737.01)","726.70 (min 681.24, max 808.97)" + "bw_file_rd-io-1mb (mb/s)","839.44 (min 808.93, max 863.71)","1022.11 (min 929.51, max 1115.86)","936.16 (min 877.35, max 1052.08)" + "bw_file_rd-o2c-1mb (mb/s)","443.48 (min 412.09, max 462.11)","551.30 (min 478.85, max 640.82)","514.92 (min 467.58, max 584.88)" + "bw_mem-bcopy-16mb (mb/s)","680.62 (min 656.28, max 700.77)","839.11 (min 760.85, max 924.59)","773.95 (min 704.44, max 876.66)" + "bw_mem-bcopy-1mb (mb/s)","720.72 (min 682.83, max 770.18)","927.90 (min 789.89, max 1060.82)","860.74 (min 767.02, max 1018.50)" + "bw_mem-bcopy-2mb (mb/s)","685.52 (min 657.14, max 733.41)","874.06 (min 757.86, max 992.56)","803.28 (min 735.74, max 919.68)" + "bw_mem-bcopy-4mb (mb/s)","702.03 (min 684.93, max 721.11)","888.30 (min 791.14, max 993.05)","786.86 (min 715.44, max 901.51)" + "bw_mem-bcopy-8mb (mb/s)","716.08 (min 691.20, max 740.33)","905.04 (min 794.20, max 1018.20)","770.79 (min 699.97, max 901.61)" + "bw_mem-bzero-16mb (mb/s)","1734.34 (min 1684.21, max 1768.93)","1940.25 (min 1748.25, max 2126.25)","1869.45 (min 1731.98, max 2129.08)" + "bw_mem-bzero-1mb (mb/s)","1223.03 (min 682.83, max 1767.41)","1427.89 (min 789.89, max 2127.28)","1360.25 (min 767.02, max 2129.55)" + "bw_mem-bzero-2mb (mb/s)","1205.38 (min 657.14, max 1767.41)","1400.07 (min 757.86, max 2124.65)","1331.83 (min 735.74, max 2129.55)" + "bw_mem-bzero-4mb (mb/s)","1215.14 (min 684.93, max 1771.22)","1409.99 (min 791.14, max 2126.91)","1323.07 (min 715.44, max 2130.30)" + "bw_mem-bzero-8mb (mb/s)","1225.86 (min 691.20, max 1781.14)","1421.45 (min 794.20, max 2124.83)","1319.03 (min 699.97, max 2129.36)" + "bw_mem-cp-16mb (mb/s)","395.42 (min 381.01, max 405.48)","508.67 (min 443.12, max 571.82)","439.42 (min 411.29, max 492.47)" + "bw_mem-cp-1mb (mb/s)","1133.11 (min 381.10, max 1932.74)","1289.61 (min 438.08, max 2283.11)","1237.66 (min 411.47, max 2268.00)" + "bw_mem-cp-2mb (mb/s)","1104.43 (min 374.60, max 1854.94)","1271.76 (min 439.13, max 2205.07)","1187.72 (min 401.53, max 2207.10)" + "bw_mem-cp-4mb (mb/s)","1093.19 (min 398.96, max 1817.08)","1249.43 (min 434.40, max 2166.46)","1175.21 (min 410.47, max 2171.16)" + "bw_mem-cp-8mb (mb/s)","1096.97 (min 413.05, max 1805.67)","1254.34 (min 479.82, max 2153.72)","1168.07 (min 416.45, max 2156.04)" + "bw_mem-fcp-16mb (mb/s)","680.05 (min 662.11, max 693.36)","816.48 (min 729.13, max 905.80)","843.99 (min 759.88, max 965.43)" + "bw_mem-fcp-1mb (mb/s)","1253.14 (min 754.72, max 1767.41)","1450.43 (min 856.60, max 2127.28)","1363.96 (min 784.81, max 2129.55)" + "bw_mem-fcp-2mb (mb/s)","1239.40 (min 720.46, max 1767.41)","1443.63 (min 838.57, max 2124.65)","1367.84 (min 795.07, max 2129.55)" + "bw_mem-fcp-4mb (mb/s)","1248.90 (min 737.60, max 1771.22)","1453.26 (min 858.65, max 2126.91)","1347.89 (min 756.86, max 2130.30)" + "bw_mem-fcp-8mb (mb/s)","1261.19 (min 762.56, max 1781.14)","1473.56 (min 882.90, max 2124.83)","1360.30 (min 773.92, max 2129.36)" + "bw_mem-frd-16mb (mb/s)","1044.45 (min 1007.81, max 1077.44)","1271.34 (min 1114.83, max 1446.00)","1170.00 (min 1098.75, max 1306.66)" + "bw_mem-frd-1mb (mb/s)","968.42 (min 754.72, max 1179.84)","1192.10 (min 856.60, max 1573.25)","1100.20 (min 784.81, max 1542.61)" + "bw_mem-frd-2mb (mb/s)","886.57 (min 720.46, max 1043.30)","1107.97 (min 838.57, max 1432.15)","1016.36 (min 795.07, max 1294.92)" + "bw_mem-frd-4mb (mb/s)","902.01 (min 737.60, max 1070.38)","1098.26 (min 858.65, max 1435.49)","1011.59 (min 756.86, max 1366.12)" + "bw_mem-frd-8mb (mb/s)","908.57 (min 762.56, max 1068.66)","1141.53 (min 882.90, max 1442.48)","1003.00 (min 773.92, max 1375.52)" + "bw_mem-fwr-16mb (mb/s)","1744.19 (min 1697.07, max 1785.91)","1948.47 (min 1750.55, max 2147.65)","1878.33 (min 1745.39, max 2138.47)" + "bw_mem-fwr-1mb (mb/s)","1510.52 (min 1125.91, max 1932.74)","1739.87 (min 1273.89, max 2283.11)","1668.01 (min 1199.86, max 2268.00)" + "bw_mem-fwr-2mb (mb/s)","1411.88 (min 981.84, max 1854.94)","1648.53 (min 1105.58, max 2205.07)","1546.88 (min 1073.73, max 2207.10)" + "bw_mem-fwr-4mb (mb/s)","1403.67 (min 968.05, max 1817.08)","1597.72 (min 1042.75, max 2166.46)","1544.82 (min 1092.60, max 2171.16)" + "bw_mem-fwr-8mb (mb/s)","1397.79 (min 990.71, max 1805.67)","1616.90 (min 1130.26, max 2153.72)","1523.88 (min 985.95, max 2156.04)" + "bw_mem-rd-16mb (mb/s)","1061.45 (min 1012.98, max 1092.75)","1296.52 (min 1154.57, max 1453.09)","1198.89 (min 1107.96, max 1370.68)" + "bw_mem-rd-1mb (mb/s)","884.55 (min 556.79, max 1234.35)","1115.27 (min 648.09, max 1650.77)","1022.60 (min 579.37, max 1591.51)" + "bw_mem-rd-2mb (mb/s)","783.16 (min 486.38, max 1073.92)","977.91 (min 568.99, max 1439.63)","892.41 (min 530.50, max 1389.61)" + "bw_mem-rd-4mb (mb/s)","813.40 (min 502.07, max 1089.32)","1023.00 (min 653.17, max 1426.02)","912.16 (min 568.83, max 1383.13)" + "bw_mem-rd-8mb (mb/s)","851.74 (min 577.95, max 1085.19)","1083.52 (min 764.31, max 1438.33)","921.44 (min 590.23, max 1383.13)" + "bw_mem-rdwr-16mb (mb/s)","672.82 (min 659.06, max 688.62)","903.05 (min 789.77, max 1031.26)","653.37 (min 586.40, max 771.05)" + "bw_mem-rdwr-1mb (mb/s)","480.16 (min 381.10, max 596.48)","615.93 (min 438.08, max 812.08)","547.06 (min 411.47, max 708.59)" + "bw_mem-rdwr-2mb (mb/s)","448.92 (min 374.60, max 515.86)","580.74 (min 439.13, max 772.80)","515.12 (min 401.53, max 702.86)" + "bw_mem-rdwr-4mb (mb/s)","489.55 (min 398.96, max 603.68)","640.10 (min 434.40, max 881.06)","520.62 (min 410.47, max 678.89)" + "bw_mem-rdwr-8mb (mb/s)","527.31 (min 413.05, max 674.02)","680.65 (min 479.82, max 924.11)","531.07 (min 416.45, max 722.09)" + "bw_mem-wr-16mb (mb/s)","694.95 (min 661.76, max 737.02)","927.61 (min 813.05, max 1038.89)","644.35 (min 583.58, max 753.15)" + "bw_mem-wr-1mb (mb/s)","562.90 (min 544.86, max 596.48)","732.36 (min 626.37, max 848.61)","638.92 (min 560.01, max 764.29)" + "bw_mem-wr-2mb (mb/s)","501.45 (min 476.53, max 547.20)","671.22 (min 559.91, max 772.80)","584.84 (min 518.87, max 702.86)" + "bw_mem-wr-4mb (mb/s)","565.99 (min 502.07, max 616.71)","763.18 (min 653.06, max 887.51)","608.59 (min 534.90, max 728.07)" + "bw_mem-wr-8mb (mb/s)","638.16 (min 577.87, max 679.69)","846.88 (min 708.97, max 1013.04)","632.04 (min 560.85, max 747.45)" + "bw_mmap_rd-mo-1mb (mb/s)","1152.80 (min 1115.24, max 1190.27)","1410.30 (min 1246.88, max 1588.42)","1348.64 (min 1240.47, max 1557.90)" + "bw_mmap_rd-o2c-1mb (mb/s)","445.75 (min 410.12, max 471.70)","553.69 (min 497.02, max 669.46)","506.34 (min 470.00, max 572.08)" + "bw_pipe (mb/s)","463.99 (min 453.23, max 474.31)","567.38 (min 491.98, max 625.68)","511.98 (min 499.61, max 534.49)" + "bw_unix (mb/s)","550.16 (min 537.95, max 572.90)","656.18 (min 589.74, max 737.01)","726.70 (min 681.24, max 808.97)" + "lat_connect (us)","73.77 (min 59.56, max 87.60)","63.98 (min 51.45, max 76.63)","75.50 (min 71.71, max 77.82)" + "lat_ctx-2-128k (us)","10.12 (min 8.90, max 12.47)","9.10 (min 8.05, max 13.66)","11.68 (min 8.45, max 16.33)" + "lat_ctx-2-256k (us)","35.67 (min 9.10, max 85.73)","32.53 (min 7.09, max 79.24)","26.36 (min 8.58, max 58.00)" + "lat_ctx-4-128k (us)","24.74 (min 9.55, max 55.08)","24.06 (min 8.02, max 57.28)","32.59 (min 9.40, max 49.60)" + "lat_ctx-4-256k (us)","12.58 (min 10.84, max 15.20)","70.36 (min 7.41, max 195.04)","120.21 (min 15.56, max 197.58)" + "lat_fs-0k (num_files)","230.17 (min 206.00, max 253.00)","259.67 (min 222.00, max 292.00)","247.67 (min 226.00, max 266.00)" + "lat_fs-10k (num_files)","97.83 (min 89.00, max 115.00)","112.83 (min 101.00, max 150.00)","104.00 (min 100.00, max 111.00)" + "lat_fs-1k (num_files)","145.00 (min 129.00, max 160.00)","166.83 (min 141.00, max 191.00)","156.67 (min 149.00, max 161.00)" + "lat_fs-4k (num_files)","142.17 (min 126.00, max 155.00)","161.83 (min 130.00, max 190.00)","149.00 (min 142.00, max 159.00)" + "lat_mem_rd-stride128-sz1000k (ns)","56.02 (min 54.49, max 58.17)","46.44 (min 41.29, max 51.32)","49.54 (min 43.10, max 52.89)" + "lat_mem_rd-stride128-sz125k (ns)","6.24 (min 6.20, max 6.29)","5.57 (min 5.55, max 5.59)","5.55 (min 5.54, max 5.56)" + "lat_mem_rd-stride128-sz250k (ns)","6.56 (min 6.53, max 6.65)","5.84 (min 5.83, max 5.84)","5.83 (min 5.83, max 5.84)" + "lat_mem_rd-stride128-sz31k (ns)","3.82 (min 2.42, max 5.23)","3.24 (min 2.16, max 4.19)","3.15 (min 2.16, max 3.65)" + "lat_mem_rd-stride128-sz50 (ns)","2.40","2.15","2.15" + "lat_mem_rd-stride128-sz500k (ns)","19.03 (min 17.22, max 21.61)","16.13 (min 12.75, max 18.51)","14.66 (min 13.14, max 16.84)" + "lat_mem_rd-stride128-sz62k (ns)","5.78 (min 5.36, max 5.87)","5.24 (min 5.22, max 5.28)","5.01 (min 4.52, max 5.26)" + "lat_mmap-1m (us)","64.33 (min 57.00, max 71.00)","57.00 (min 50.00, max 62.00)","56.33 (min 48.00, max 63.00)" + "lat_ops-double-add (ns)","3.21","2.86","2.86" + "lat_ops-double-div (ns)","17.65 (min 17.63, max 17.66)","15.75 (min 15.73, max 15.76)","15.74 (min 15.74, max 15.75)" + "lat_ops-double-mul (ns)","3.21","2.86","2.86" + "lat_ops-float-add (ns)","3.21","2.86","2.86" + "lat_ops-float-div (ns)","10.42 (min 10.42, max 10.43)","9.30 (min 9.30, max 9.31)","9.30" + "lat_ops-float-mul (ns)","3.21","2.86","2.86" + "lat_ops-int-add (ns)","0.80","0.72","0.72" + "lat_ops-int-bit (ns)","0.53","0.48","0.48" + "lat_ops-int-div (ns)","4.81","4.29 (min 4.29, max 4.30)","4.30 (min 4.29, max 4.30)" + "lat_ops-int-mod (ns)","5.08","4.53","4.54 (min 4.53, max 4.54)" + "lat_ops-int-mul (ns)","3.46 (min 3.40, max 3.49)","3.09 (min 3.04, max 3.12)","3.12 (min 3.11, max 3.12)" + "lat_ops-int64-add (ns)","0.80","0.72","0.72" + "lat_ops-int64-bit (ns)","0.53 (min 0.53, max 0.54)","0.48","0.48" + "lat_ops-int64-div (ns)","7.62 (min 7.61, max 7.63)","6.80 (min 6.79, max 6.80)","6.80" + "lat_ops-int64-mod (ns)","5.88 (min 5.88, max 5.89)","5.25","5.25 (min 5.24, max 5.25)" + "lat_ops-int64-mul (ns)","3.99 (min 3.97, max 4.07)","3.56 (min 3.54, max 3.64)","3.56 (min 3.54, max 3.59)" + "lat_pagefault (us)","0.68 (min 0.66, max 0.70)","0.58 (min 0.52, max 0.63)","1.69 (min 1.53, max 1.78)" + "lat_pipe (us)","24.07 (min 23.56, max 24.62)","21.52 (min 21.04, max 22.12)","21.79 (min 21.03, max 22.35)" + "lat_proc-exec (us)","1130.37 (min 1095.20, max 1173.20)","936.72 (min 840.17, max 1059.33)","1003.69 (min 866.86, max 1074.80)" + "lat_proc-fork (us)","1104.61 (min 926.33, max 1739.33)","818.28 (min 731.25, max 906.50)","855.18 (min 733.71, max 943.67)" "lat_proc-proccall (us)","0.01","0.01","0.01" - "lat_select (us)","38.12 (min 37.94, max 38.28)","34.11 (min 33.75, max 34.40)","38.13 (min 37.96, max 38.48)" - "lat_sem (us)","3.53 (min 2.89, max 4.27)","3.09 (min 2.59, max 3.54)","3.36 (min 2.96, max 3.84)" - "lat_sig-catch (us)","6.23 (min 6.03, max 6.45)","5.52 (min 5.27, max 5.68)","6.18 (min 5.95, max 6.37)" - "lat_sig-install (us)","0.74 (min 0.72, max 0.79)","0.66 (min 0.65, max 0.70)","0.75 (min 0.72, max 0.79)" - "lat_sig-prot (us)","0.71 (min 0.55, max 0.93)","0.65 (min 0.48, max 0.80)","0.83 (min 0.58, max 1.48)" - "lat_syscall-fstat (us)","2.21 (min 2.14, max 2.32)","1.98 (min 1.89, max 2.10)","2.22 (min 2.12, max 2.32)" - "lat_syscall-null (us)","0.51 (min 0.51, max 0.52)","0.46 (min 0.46, max 0.50)","0.52 (min 0.51, max 0.56)" - "lat_syscall-open (us)","214.57 (min 153.03, max 305.39)","175.20 (min 155.95, max 202.88)","174.43 (min 153.94, max 210.15)" - "lat_syscall-read (us)","0.92 (min 0.90, max 0.99)","0.82 (min 0.80, max 0.88)","0.92 (min 0.90, max 0.99)" - "lat_syscall-stat (us)","5.37 (min 5.23, max 5.53)","4.80 (min 4.66, max 5.01)","5.39 (min 5.25, max 5.60)" - "lat_syscall-write (us)","0.86 (min 0.84, max 0.89)","0.78 (min 0.75, max 0.84)","0.87 (min 0.84, max 0.94)" - "lat_tcp (us)","1.02","0.92 (min 0.91, max 0.97)","1.03 (min 1.02, max 1.09)" - "lat_unix (us)","33.40 (min 32.86, max 34.65)","29.79 (min 28.64, max 30.84)","33.61 (min 33.07, max 34.90)" - "latency_for_0.50_mb_block_size (nanosec)","17.11 (min 11.46, max 21.66)","17.27 (min 13.44, max 20.26)","15.35 (min 10.09, max 21.82)" - "latency_for_1.00_mb_block_size (nanosec)","28.48 (min 0.00, max 58.01)","25.22 (min 0.00, max 52.51)","24.41 (min 0.00, max 54.17)" - "pipe_bandwidth (mb\s)","443.15 (min 435.10, max 453.94)","515.70 (min 492.51, max 600.38)","498.69 (min 451.80, max 545.64)" - "pipe_latency (microsec)","28.82 (min 28.37, max 29.43)","25.77 (min 25.50, max 26.22)","28.93 (min 28.25, max 29.53)" + "lat_select (us)","35.44 (min 35.10, max 35.63)","31.58 (min 31.34, max 32.00)","31.59 (min 31.47, max 31.79)" + "lat_sem (us)","3.05 (min 2.72, max 3.89)","2.55 (min 2.34, max 2.82)","2.92 (min 2.87, max 2.97)" + "lat_sig-catch (us)","6.27 (min 6.11, max 6.44)","5.60 (min 5.46, max 5.72)","5.55 (min 5.48, max 5.67)" + "lat_sig-install (us)","0.73 (min 0.65, max 0.80)","0.65 (min 0.58, max 0.72)","0.60 (min 0.58, max 0.63)" + "lat_sig-prot (us)","0.96 (min 0.83, max 1.10)","0.85 (min 0.75, max 0.92)","0.81 (min 0.76, max 0.86)" + "lat_syscall-fstat (us)","2.04 (min 1.97, max 2.12)","1.81 (min 1.76, max 1.89)","1.80 (min 1.74, max 1.88)" + "lat_syscall-null (us)","0.49 (min 0.41, max 0.57)","0.44 (min 0.37, max 0.51)","0.37" + "lat_syscall-open (us)","222.13 (min 175.60, max 296.65)","181.80 (min 156.61, max 200.77)","169.27 (min 160.19, max 184.77)" + "lat_syscall-read (us)","0.84 (min 0.82, max 0.89)","0.75 (min 0.73, max 0.79)","0.74 (min 0.73, max 0.74)" + "lat_syscall-stat (us)","4.64 (min 4.46, max 4.76)","4.16 (min 3.99, max 4.28)","4.24 (min 4.23, max 4.25)" + "lat_syscall-write (us)","0.79 (min 0.75, max 0.84)","0.70 (min 0.67, max 0.74)","0.67" + "lat_tcp (us)","1.00 (min 0.85, max 1.14)","0.89 (min 0.76, max 1.02)","0.76" + "lat_unix (us)","29.86 (min 25.84, max 32.02)","25.62 (min 23.44, max 27.72)","27.27 (min 26.72, max 27.91)" + "latency_for_0.50_mb_block_size (nanosec)","19.03 (min 17.22, max 21.61)","16.13 (min 12.75, max 18.51)","14.66 (min 13.14, max 16.84)" + "latency_for_1.00_mb_block_size (nanosec)","28.01 (min 0.00, max 58.17)","23.22 (min 0.00, max 51.32)","24.77 (min 0.00, max 52.89)" + "pipe_bandwidth (mb\s)","463.99 (min 453.23, max 474.31)","567.38 (min 491.98, max 625.68)","511.98 (min 499.61, max 534.49)" + "pipe_latency (microsec)","24.07 (min 23.56, max 24.62)","21.52 (min 21.04, max 22.12)","21.79 (min 21.03, max 22.35)" "procedure_call (microsec)","0.01","0.01","0.01" - "select_on_200_tcp_fds (microsec)","38.12 (min 37.94, max 38.28)","34.11 (min 33.75, max 34.40)","38.13 (min 37.96, max 38.48)" - "semaphore_latency (microsec)","3.53 (min 2.89, max 4.27)","3.09 (min 2.59, max 3.54)","3.36 (min 2.96, max 3.84)" - "signal_handler_latency (microsec)","0.74 (min 0.72, max 0.79)","0.66 (min 0.65, max 0.70)","0.75 (min 0.72, max 0.79)" - "signal_handler_overhead (microsec)","6.23 (min 6.03, max 6.45)","5.52 (min 5.27, max 5.68)","6.18 (min 5.95, max 6.37)" - "tcp_ip_connection_cost_to_localhost (microsec)","66.67 (min 66.20, max 67.85)","59.11 (min 58.18, max 59.76)","65.67 (min 63.87, max 69.00)" - "tcp_latency_using_localhost (microsec)","1.02","0.92 (min 0.91, max 0.97)","1.03 (min 1.02, max 1.09)" + "select_on_200_tcp_fds (microsec)","35.44 (min 35.10, max 35.63)","31.58 (min 31.34, max 32.00)","31.59 (min 31.47, max 31.79)" + "semaphore_latency (microsec)","3.05 (min 2.72, max 3.89)","2.55 (min 2.34, max 2.82)","2.92 (min 2.87, max 2.97)" + "signal_handler_latency (microsec)","0.73 (min 0.65, max 0.80)","0.65 (min 0.58, max 0.72)","0.60 (min 0.58, max 0.63)" + "signal_handler_overhead (microsec)","6.27 (min 6.11, max 6.44)","5.60 (min 5.46, max 5.72)","5.55 (min 5.48, max 5.67)" + "tcp_ip_connection_cost_to_localhost (microsec)","73.77 (min 59.56, max 87.60)","63.98 (min 51.45, max 76.63)","75.50 (min 71.71, max 77.82)" + "tcp_latency_using_localhost (microsec)","1.00 (min 0.85, max 1.14)","0.89 (min 0.76, max 1.02)","0.76" Dhrystone -^^^^^^^^^ +========= + Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. @@ -200,12 +207,13 @@ Execute the benchmark with the following: .. csv-table:: Dhrystone Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "cpu_clock (mhz)","1250.00","1400.00","1250.00" - "dhrystone_per_mhz (dmips/mhz)","2.84 (min 2.80, max 2.90)","2.90","2.86 (min 2.80, max 2.90)" - "dhrystone_per_second (dhrystonep)","6336405.57 (min 6250000.00, max 6451613.00)","7142857.00","6376008.13 (min 6250000.00, max 6451613.00)" + "cpu_clock (mhz)","1250.00","1400.00","1400.00" + "dhrystone_per_mhz (dmips/mhz)","2.76 (min 2.70, max 2.80)","2.75 (min 2.60, max 2.90)","2.77 (min 2.70, max 2.80)" + "dhrystone_per_second (dhrystonep)","6027183.60 (min 5882353.00, max 6250000.00)","6789289.58 (min 6451613.00, max 7142857.00)","6819923.17 (min 6666666.50, max 6896551.50)" Whetstone -^^^^^^^^^ +========= + Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -217,20 +225,22 @@ Execute the benchmark with the following: .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "whetstone (mips)","5000.00","5000.00","5000.00" + "whetstone (mips)","4444.43 (min 3333.30, max 5000.00)","5000.00","5000.00" Linpack -^^^^^^^ +======= + Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "linpack (kflops)","513733.00 (min 511865.00, max 515438.00)","577158.63 (min 576063.00, max 577662.00)","515214.00 (min 513050.00, max 516718.00)" + "linpack (kflops)","515140.50 (min 508416.00, max 518513.00)","581699.00 (min 581477.00, max 581921.00)","578855.50 (min 578148.00, max 579563.00)" NBench -^^^^^^ +====== + NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at @@ -240,18 +250,19 @@ https://nbench.io/articles/index.html .. csv-table:: NBench Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "assignment (iterations)","12.92 (min 12.80, max 12.95)","14.49 (min 14.44, max 14.53)","12.93 (min 12.85, max 12.96)" - "fourier (iterations)","20380.88 (min 20371.00, max 20385.00)","22828.75 (min 22827.00, max 22831.00)","20383.88 (min 20381.00, max 20385.00)" - "fp_emulation (iterations)","192.50 (min 192.47, max 192.52)","215.61 (min 215.57, max 215.65)","192.52 (min 192.49, max 192.54)" - "huffman (iterations)","1057.11 (min 1056.50, max 1057.40)","1183.90 (min 1183.30, max 1184.30)","1057.24 (min 1056.70, max 1057.40)" - "idea (iterations)","3075.31 (min 3074.90, max 3075.60)","3444.45 (min 3444.20, max 3444.80)","3075.58 (min 3075.30, max 3075.70)" - "lu_decomposition (iterations)","472.14 (min 469.83, max 473.70)","527.41 (min 522.26, max 529.95)","472.13 (min 469.78, max 473.74)" - "neural_net (iterations)","7.73 (min 7.72, max 7.73)","8.65 (min 8.64, max 8.66)","7.72 (min 7.71, max 7.73)" - "numeric_sort (iterations)","559.47 (min 553.45, max 562.47)","625.06 (min 616.37, max 629.09)","559.74 (min 550.44, max 561.86)" - "string_sort (iterations)","146.35 (min 146.33, max 146.37)","163.92 (min 163.91, max 163.94)","146.36 (min 146.35, max 146.37)" + "assignment (iterations)","12.63 (min 12.55, max 12.73)","14.16 (min 14.15, max 14.17)","14.14 (min 14.12, max 14.15)" + "fourier (iterations)","18561.17 (min 18558.00, max 18567.00)","20791.00 (min 20785.00, max 20798.00)","20792.25 (min 20790.00, max 20796.00)" + "fp_emulation (iterations)","182.44 (min 182.30, max 182.54)","204.36 (min 204.17, max 204.44)","204.39 (min 204.37, max 204.41)" + "huffman (iterations)","1058.38 (min 1049.10, max 1063.20)","1187.26 (min 1174.70, max 1191.10)","1190.58 (min 1190.00, max 1190.90)" + "idea (iterations)","2738.27 (min 2737.60, max 2738.60)","3066.96 (min 3066.60, max 3067.30)","3067.23 (min 3067.00, max 3067.40)" + "lu_decomposition (iterations)","472.09 (min 469.76, max 474.80)","530.93 (min 526.72, max 532.68)","529.40 (min 526.80, max 532.45)" + "neural_net (iterations)","8.60 (min 8.51, max 8.64)","9.65 (min 9.53, max 9.68)","9.68 (min 9.67, max 9.68)" + "numeric_sort (iterations)","482.67 (min 477.04, max 484.80)","538.40 (min 533.26, max 542.91)","537.80 (min 527.29, max 542.82)" + "string_sort (iterations)","150.20 (min 150.14, max 150.27)","168.20 (min 168.14, max 168.29)","168.19 (min 168.17, max 168.21)" Stream -^^^^^^ +====== + STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -271,13 +282,14 @@ Execute the benchmark with the following: .. csv-table:: Stream Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "add (mb/s)","1403.63 (min 1363.30, max 1452.60)","1628.69 (min 1443.40, max 1805.40)","1690.80 (min 1688.10, max 1693.80)" - "copy (mb/s)","1471.36 (min 1421.90, max 1537.40)","1877.63 (min 1635.60, max 2125.00)","1869.34 (min 1836.10, max 1926.50)" - "scale (mb/s)","1605.57 (min 1561.70, max 1663.20)","2031.11 (min 1793.40, max 2270.80)","1937.30 (min 1923.80, max 1949.60)" - "triad (mb/s)","1378.53 (min 1343.40, max 1423.00)","1676.24 (min 1479.30, max 1872.10)","1656.76 (min 1651.70, max 1662.10)" + "add (mb/s)","1376.20 (min 1329.10, max 1411.30)","1580.70 (min 1463.60, max 1811.00)","1704.70" + "copy (mb/s)","1462.62 (min 1366.90, max 1515.60)","1838.92 (min 1650.20, max 2245.20)","1876.80" + "scale (mb/s)","1565.95 (min 1474.00, max 1610.80)","1948.28 (min 1786.10, max 2285.60)","1921.20" + "triad (mb/s)","1349.88 (min 1303.10, max 1385.40)","1615.13 (min 1494.30, max 1856.90)","1667.70" CoreMarkPro -^^^^^^^^^^^ +=========== + CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, @@ -287,33 +299,34 @@ and floating-point workloads, and data sets for utilizing larger memory subsyste .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "cjpeg-rose7-preset (workloads/)","37.30 (min 37.17, max 37.45)","41.84 (min 41.67, max 42.02)","37.38 (min 36.90, max 37.59)" - "core (workloads/)","0.27","0.30","0.27" - "coremark-pro ()","784.22 (min 761.93, max 799.55)","874.96 (min 849.21, max 918.16)","800.81 (min 770.03, max 828.46)" - "linear_alg-mid-100x100-sp (workloads/)","13.10 (min 13.09, max 13.11)","14.68 (min 14.66, max 14.70)","13.10 (min 13.09, max 13.11)" - "loops-all-mid-10k-sp (workloads/)","0.59 (min 0.59, max 0.60)","0.67 (min 0.66, max 0.70)","0.62 (min 0.60, max 0.63)" - "nnet_test (workloads/)","0.97","1.09 (min 1.08, max 1.09)","0.97" - "parser-125k (workloads/)","7.38 (min 7.30, max 7.46)","8.30 (min 8.20, max 8.62)","7.64 (min 7.41, max 7.75)" - "radix2-big-64k (workloads/)","42.96 (min 33.97, max 51.13)","46.53 (min 35.36, max 63.36)","46.38 (min 36.20, max 60.27)" - "sha-test (workloads/)","71.81 (min 70.92, max 71.94)","80.56 (min 80.00, max 80.65)","72.14 (min 71.43, max 72.46)" - "zip-test (workloads/)","18.52","20.84 (min 20.41, max 21.28)","18.96 (min 18.52, max 19.23)" + "cjpeg-rose7-preset (workloads/)","37.27 (min 37.04, max 37.59)","41.73 (min 41.49, max 41.84)","41.84 (min 41.67, max 42.02)" + "core (workloads/)","0.27","0.30","0.30" + "coremark-pro ()","793.57 (min 764.61, max 819.06)","888.50 (min 846.80, max 928.38)","879.40 (min 859.11, max 912.05)" + "linear_alg-mid-100x100-sp (workloads/)","13.09 (min 13.08, max 13.11)","14.67 (min 14.66, max 14.70)","14.68 (min 14.68, max 14.69)" + "loops-all-mid-10k-sp (workloads/)","0.60 (min 0.60, max 0.62)","0.67 (min 0.67, max 0.70)","0.68 (min 0.66, max 0.70)" + "nnet_test (workloads/)","0.96 (min 0.95, max 0.96)","1.08","1.08 (min 1.07, max 1.08)" + "parser-125k (workloads/)","7.65 (min 7.41, max 8.00)","8.52 (min 8.13, max 8.85)","8.68 (min 8.26, max 9.01)" + "radix2-big-64k (workloads/)","44.85 (min 33.86, max 54.28)","51.45 (min 34.28, max 69.88)","44.89 (min 39.02, max 55.47)" + "sha-test (workloads/)","71.77 (min 70.92, max 72.46)","80.12 (min 78.74, max 81.30)","80.86 (min 80.65, max 81.30)" + "zip-test (workloads/)","19.07 (min 18.52, max 20.00)","21.37 (min 20.83, max 22.22)","21.28 (min 20.83, max 21.74)" .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "cjpeg-rose7-preset (workloads/)","73.92 (min 73.53, max 74.07)","82.82 (min 81.97, max 83.33)","74.28 (min 73.53, max 74.63)" - "core (workloads/)","0.54","0.60","0.54" - "coremark-pro ()","1335.06 (min 1324.28, max 1347.82)","1527.89 (min 1465.32, max 1576.42)","1407.68 (min 1356.48, max 1438.96)" - "linear_alg-mid-100x100-sp (workloads/)","26.18 (min 26.16, max 26.19)","29.33 (min 29.31, max 29.36)","26.19 (min 26.12, max 26.21)" - "loops-all-mid-10k-sp (workloads/)","1.03 (min 1.02, max 1.05)","1.19 (min 1.14, max 1.24)","1.11 (min 1.04, max 1.13)" - "nnet_test (workloads/)","1.94 (min 1.93, max 1.94)","2.17","1.94 (min 1.93, max 1.94)" - "parser-125k (workloads/)","10.55 (min 10.26, max 10.93)","12.28 (min 11.05, max 13.42)","11.24 (min 10.00, max 12.12)" - "radix2-big-64k (workloads/)","35.17 (min 33.06, max 38.93)","44.61 (min 36.64, max 51.47)","47.05 (min 40.80, max 50.65)" - "sha-test (workloads/)","142.86","160.01 (min 158.73, max 161.29)","143.89 (min 142.86, max 144.93)" - "zip-test (workloads/)","34.49 (min 33.90, max 35.09)","39.14 (min 37.74, max 40.00)","35.96 (min 35.09, max 36.36)" + "cjpeg-rose7-preset (workloads/)","73.38 (min 71.94, max 74.07)","82.31 (min 81.30, max 83.33)","83.10 (min 82.64, max 83.33)" + "core (workloads/)","0.54","0.60","0.60" + "coremark-pro ()","1356.04 (min 1322.58, max 1417.19)","1525.15 (min 1472.72, max 1594.34)","1560.33 (min 1506.68, max 1589.98)" + "linear_alg-mid-100x100-sp (workloads/)","26.15 (min 26.12, max 26.18)","29.32 (min 29.29, max 29.34)","29.33 (min 29.33, max 29.34)" + "loops-all-mid-10k-sp (workloads/)","1.05 (min 1.02, max 1.12)","1.18 (min 1.14, max 1.26)","1.21 (min 1.15, max 1.25)" + "nnet_test (workloads/)","1.92","2.15","2.15" + "parser-125k (workloads/)","12.19 (min 10.81, max 13.42)","13.68 (min 11.63, max 16.00)","14.12 (min 12.90, max 15.38)" + "radix2-big-64k (workloads/)","35.34 (min 32.03, max 46.62)","40.11 (min 36.05, max 45.74)","44.46 (min 40.00, max 49.31)" + "sha-test (workloads/)","140.33 (min 136.99, max 144.93)","156.69 (min 153.85, max 158.73)","160.44 (min 158.73, max 161.29)" + "zip-test (workloads/)","35.29 (min 33.90, max 36.36)","40.31 (min 38.46, max 42.55)","40.60 (min 38.46, max 41.67)" MultiBench -^^^^^^^^^^ +========== + MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -332,52 +345,54 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" - "4m-check (workloads/)","274.66 (min 267.24, max 283.06)","342.64 (min 302.01, max 368.24)","302.78 (min 279.33, max 340.88)" - "4m-check-reassembly (workloads/)","58.33 (min 56.63, max 60.42)","72.79 (min 62.42, max 80.13)","63.46 (min 56.21, max 74.35)" - "4m-check-reassembly-tcp (workloads/)","37.86 (min 36.82, max 38.76)","46.33 (min 41.32, max 49.80)","41.02 (min 37.48, max 46.04)" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","21.83 (min 21.34, max 22.42)","27.22 (min 23.86, max 29.30)","24.15 (min 22.35, max 27.04)" - "4m-check-reassembly-tcp-x264w2 (workloads/)","1.59 (min 1.58, max 1.61)","1.80 (min 1.76, max 1.84)","1.62 (min 1.59, max 1.66)" - "4m-cmykw2 (workloads/)","178.88 (min 173.61, max 184.50)","220.64 (min 194.55, max 242.13)","192.61 (min 172.86, max 216.45)" - "4m-cmykw2-rotatew2 (workloads/)","35.01 (min 34.05, max 35.93)","44.14 (min 38.78, max 47.58)","37.56 (min 33.82, max 41.81)" - "4m-reassembly (workloads/)","46.17 (min 43.86, max 51.57)","57.21 (min 49.58, max 61.27)","50.47 (min 45.79, max 57.44)" - "4m-rotatew2 (workloads/)","40.02 (min 39.09, max 41.03)","49.21 (min 44.25, max 52.47)","43.68 (min 40.95, max 47.92)" - "4m-tcp-mixed (workloads/)","93.04 (min 91.43, max 95.24)","110.80 (min 103.90, max 115.11)","97.63 (min 93.02, max 104.58)" - "4m-x264w2 (workloads/)","1.63 (min 1.56, max 1.65)","1.87 (min 1.81, max 1.91)","1.65 (min 1.58, max 1.70)" + "4m-check (workloads/)","272.76 (min 272.01, max 273.58)","320.84 (min 305.89, max 373.86)","332.22 (min 303.95, max 360.49)" + "4m-check-reassembly (workloads/)","56.07 (min 55.77, max 56.34)","72.48 (min 62.27, max 86.51)","75.46 (min 73.05, max 77.88)" + "4m-check-reassembly-tcp (workloads/)","37.54 (min 37.43, max 37.65)","45.68 (min 41.81, max 50.00)","48.46 (min 47.89, max 49.02)" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","21.53 (min 21.34, max 21.79)","25.33 (min 24.19, max 29.50)","26.49 (min 23.87, max 29.11)" + "4m-check-reassembly-tcp-x264w2 (workloads/)","1.57 (min 1.57, max 1.58)","1.77 (min 1.74, max 1.83)","1.71 (min 1.71, max 1.72)" + "4m-cmykw2 (workloads/)","176.42 (min 175.59, max 176.99)","205.20 (min 197.04, max 236.97)","217.95 (min 196.66, max 239.23)" + "4m-cmykw2-rotatew2 (workloads/)","32.81 (min 32.66, max 33.09)","39.43 (min 37.11, max 45.21)","42.08 (min 38.12, max 46.05)" + "4m-reassembly (workloads/)","43.05 (min 42.50, max 43.90)","61.98 (min 47.62, max 85.25)","68.29 (min 57.97, max 78.62)" + "4m-rotatew2 (workloads/)","38.84 (min 38.55, max 39.03)","45.22 (min 42.18, max 51.87)","47.93 (min 44.39, max 51.47)" + "4m-tcp-mixed (workloads/)","98.36 (min 97.56, max 98.77)","113.23 (min 111.11, max 123.08)","116.62 (min 111.11, max 122.14)" + "4m-x264w2 (workloads/)","1.58 (min 1.52, max 1.63)","1.82 (min 1.72, max 1.91)","1.84 (min 1.82, max 1.86)" "empty-wld (workloads/)","1.00","1.00","1.00" - "idct-4m (workloads/)","16.71 (min 16.49, max 16.78)","18.89 (min 18.56, max 19.11)","16.89 (min 16.78, max 17.10)" - "idct-4mw1 (workloads/)","16.71 (min 16.49, max 16.79)","18.89 (min 18.56, max 19.11)","16.91 (min 16.76, max 17.10)" - "ippktcheck-4m (workloads/)","274.52 (min 265.53, max 283.67)","340.93 (min 303.62, max 364.38)","302.69 (min 280.36, max 339.58)" - "ippktcheck-4mw1 (workloads/)","274.24 (min 267.87, max 282.61)","342.81 (min 303.21, max 368.08)","302.30 (min 278.27, max 338.89)" - "ipres-4m (workloads/)","58.88 (min 56.58, max 61.63)","73.72 (min 63.94, max 79.79)","66.26 (min 60.61, max 74.85)" - "ipres-4mw1 (workloads/)","58.64 (min 56.97, max 60.66)","73.94 (min 64.16, max 80.30)","66.06 (min 60.68, max 74.55)" - "md5-4m (workloads/)","21.83 (min 21.29, max 22.23)","25.41 (min 24.13, max 26.17)","22.62 (min 21.41, max 23.98)" - "md5-4mw1 (workloads/)","22.12 (min 21.82, max 22.44)","25.67 (min 24.51, max 26.52)","22.83 (min 22.00, max 24.08)" - "rgbcmyk-4m (workloads/)","56.75 (min 56.35, max 57.18)","64.33 (min 63.35, max 65.02)","56.81 (min 56.58, max 57.05)" - "rgbcmyk-4mw1 (workloads/)","56.65 (min 56.37, max 57.14)","64.40 (min 63.33, max 65.08)","56.79 (min 56.66, max 57.05)" - "rotate-4ms1 (workloads/)","16.66 (min 16.20, max 17.15)","20.43 (min 18.28, max 21.82)","18.07 (min 16.78, max 20.02)" - "rotate-4ms1w1 (workloads/)","16.68 (min 16.31, max 17.16)","20.58 (min 18.29, max 22.77)","18.09 (min 16.95, max 20.03)" - "rotate-4ms64 (workloads/)","16.83 (min 16.47, max 17.25)","20.60 (min 18.40, max 22.02)","18.26 (min 17.09, max 20.15)" - "rotate-4ms64w1 (workloads/)","16.82 (min 16.46, max 17.25)","20.76 (min 18.34, max 23.08)","18.28 (min 17.10, max 20.20)" - "x264-4mq (workloads/)","0.50","0.57 (min 0.56, max 0.57)","0.50 (min 0.50, max 0.51)" - "x264-4mqw1 (workloads/)","0.50","0.56 (min 0.56, max 0.57)","0.50 (min 0.50, max 0.51)" + "idct-4m (workloads/)","16.64 (min 16.60, max 16.66)","18.99 (min 18.49, max 19.69)","19.33 (min 19.05, max 19.61)" + "idct-4mw1 (workloads/)","16.65 (min 16.64, max 16.66)","18.99 (min 18.49, max 19.69)","19.34 (min 19.05, max 19.62)" + "ippktcheck-4m (workloads/)","270.95 (min 270.27, max 271.62)","320.65 (min 303.51, max 369.22)","334.19 (min 305.44, max 362.95)" + "ippktcheck-4mw1 (workloads/)","273.45 (min 272.78, max 274.42)","321.05 (min 306.26, max 372.58)","334.33 (min 306.19, max 362.48)" + "ipres-4m (workloads/)","53.57 (min 53.42, max 53.71)","64.68 (min 60.24, max 74.44)","67.79 (min 62.66, max 72.92)" + "ipres-4mw1 (workloads/)","53.94 (min 53.51, max 54.27)","64.58 (min 60.44, max 74.93)","67.68 (min 62.55, max 72.82)" + "md5-4m (workloads/)","22.10 (min 21.90, max 22.20)","25.32 (min 24.70, max 26.60)","26.04 (min 25.27, max 26.80)" + "md5-4mw1 (workloads/)","22.29 (min 22.23, max 22.34)","25.25 (min 24.69, max 26.81)","25.90 (min 24.91, max 26.89)" + "rgbcmyk-4m (workloads/)","56.14 (min 56.04, max 56.32)","63.75 (min 63.01, max 64.89)","63.24 (min 61.79, max 64.70)" + "rgbcmyk-4mw1 (workloads/)","56.08 (min 55.96, max 56.29)","63.86 (min 63.07, max 64.91)","64.11 (min 63.47, max 64.75)" + "rotate-4ms1 (workloads/)","16.34 (min 16.32, max 16.35)","19.50 (min 18.34, max 21.71)","19.82 (min 18.13, max 21.51)" + "rotate-4ms1w1 (workloads/)","16.39 (min 16.37, max 16.41)","19.44 (min 18.23, max 21.62)","19.79 (min 18.18, max 21.40)" + "rotate-4ms64 (workloads/)","16.53 (min 16.52, max 16.56)","19.61 (min 18.38, max 21.89)","20.05 (min 18.32, max 21.78)" + "rotate-4ms64w1 (workloads/)","16.48 (min 16.46, max 16.50)","19.68 (min 18.38, max 22.23)","20.04 (min 18.34, max 21.75)" + "x264-4mq (workloads/)","0.49 (min 0.48, max 0.50)","0.56 (min 0.55, max 0.57)","0.56 (min 0.56, max 0.57)" + "x264-4mqw1 (workloads/)","0.49 (min 0.49, max 0.50)","0.56 (min 0.54, max 0.57)","0.56" + +| Boot-time Measurement ---------------------- +===================== Boot media: MMCSD -^^^^^^^^^^^^^^^^^ +----------------- .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am62xx_lp_sk-fs: Boot time in seconds: avg(min,max)","am62xx_sk-fs: Boot time in seconds: avg(min,max)","am62xxsip_sk-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","17.36 (min 15.98, max 25.46)","16.20 (min 14.74, max 22.20)","15.91 (min 14.98, max 25.80)" + "Linux boot time from SD with default rootfs (20 boot cycles)","18.28 (min 17.43, max 19.66)","18.15 (min 17.32, max 36.26)","18.15 (min 17.28, max 20.55)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. | -ALSA SoC Audio Driver ---------------------- +ALSA SoC Audio +============== #. Access type - RW\_INTERLEAVED #. Channels - 2 @@ -387,63 +402,64 @@ ALSA SoC Audio Driver .. csv-table:: Audio Capture :header: "Sampling Rate (Hz)","am62xx_lp_sk-fs: Throughput (bits/sec)","am62xx_lp_sk-fs: CPU Load (%)","am62xx_sk-fs: Throughput (bits/sec)","am62xx_sk-fs: CPU Load (%)","am62xxsip_sk-fs: Throughput (bits/sec)","am62xxsip_sk-fs: CPU Load (%)" - "11025","352792.79 (min 352756.00, max 352798.00)","0.19 (min 0.13, max 0.67)","352798.88 (min 352798.00, max 352801.00)","0.13 (min 0.11, max 0.19)","352798.25 (min 352797.00, max 352800.00)","0.14 (min 0.11, max 0.18)" - "16000","511993.07 (min 511964.00, max 511998.00)","0.19 (min 0.11, max 0.55)","511999.38 (min 511997.00, max 512002.00)","0.11 (min 0.09, max 0.17)","511998.17 (min 511996.00, max 512002.00)","0.17 (min 0.08, max 0.33)" - "22050","705595.14 (min 705591.00, max 705613.00)","0.19 (min 0.16, max 0.24)","705594.50 (min 705591.00, max 705597.00)","0.16 (min 0.15, max 0.21)","705594.42 (min 705589.00, max 705600.00)","0.17 (min 0.14, max 0.22)" - "24000","705593.71 (min 705585.00, max 705596.00)","0.21 (min 0.18, max 0.28)","705596.75 (min 705595.00, max 705599.00)","0.19 (min 0.16, max 0.25)","705596.33 (min 705594.00, max 705600.00)","0.19 (min 0.14, max 0.25)" - "32000","1023989.36 (min 1023960.00, max 1023994.00)","0.14 (min 0.10, max 0.26)","1023996.25 (min 1023993.00, max 1023999.00)","0.12 (min 0.10, max 0.17)","1023995.50 (min 1023992.00, max 1024002.00)","0.18 (min 0.09, max 0.75)" - "44100","1411184.36 (min 1411137.00, max 1411193.00)","0.29 (min 0.26, max 0.35)","1411195.00 (min 1411191.00, max 1411199.00)","0.26 (min 0.23, max 0.31)","1411193.67 (min 1411188.00, max 1411203.00)","0.26 (min 0.23, max 0.32)" - "48000","1535982.00 (min 1535927.00, max 1535992.00)","0.30 (min 0.11, max 1.33)","1535994.50 (min 1535990.00, max 1535999.00)","0.15 (min 0.10, max 0.26)","1535993.17 (min 1535987.00, max 1536003.00)","0.27 (min 0.10, max 0.92)" - "88200","2822360.21 (min 2822256.00, max 2822378.00)","0.52 (min 0.45, max 0.59)","2822385.38 (min 2822376.00, max 2822395.00)","0.49 (min 0.44, max 0.59)","2822380.33 (min 2822369.00, max 2822401.00)","0.48 (min 0.38, max 0.56)" - "96000","3071963.64 (min 3071875.00, max 3071978.00)","0.25 (min 0.17, max 0.50)","3071983.00 (min 3071971.00, max 3071992.00)","0.24 (min 0.19, max 0.39)","3071976.00 (min 3071954.00, max 3071998.00)","0.32 (min 0.16, max 1.26)" + "11025","352797.40 (min 352797.00, max 352798.00)","0.20 (min 0.17, max 0.31)","352798.20 (min 352797.00, max 352799.00)","0.22 (min 0.13, max 0.32)","352799.67 (min 352799.00, max 352800.00)","0.30 (min 0.25, max 0.35)" + "16000","511997.20 (min 511996.00, max 511998.00)","0.14 (min 0.12, max 0.15)","511998.00 (min 511997.00, max 512000.00)","0.15 (min 0.09, max 0.28)","512001.00 (min 511999.00, max 512002.00)","0.14 (min 0.13, max 0.14)" + "22050","705591.80 (min 705591.00, max 705593.00)","0.18 (min 0.17, max 0.21)","705594.00 (min 705592.00, max 705596.00)","0.17 (min 0.15, max 0.22)","705593.67 (min 705593.00, max 705595.00)","0.19 (min 0.18, max 0.20)" + "24000","705594.40 (min 705593.00, max 705595.00)","0.20 (min 0.19, max 0.23)","705596.00 (min 705594.00, max 705598.00)","0.19 (min 0.16, max 0.24)","705598.33 (min 705597.00, max 705599.00)","0.19 (min 0.18, max 0.20)" + "32000","1023992.60 (min 1023991.00, max 1023994.00)","0.12 (min 0.11, max 0.13)","1023995.00 (min 1023992.00, max 1023998.00)","0.28 (min 0.09, max 0.97)","1024000.00 (min 1023997.00, max 1024002.00)","0.13 (min 0.12, max 0.14)" + "44100","1411189.80 (min 1411188.00, max 1411192.00)","0.28 (min 0.27, max 0.30)","1411192.80 (min 1411189.00, max 1411197.00)","0.26 (min 0.23, max 0.30)","1411200.33 (min 1411196.00, max 1411203.00)","0.27" + "48000","1535988.80 (min 1535986.00, max 1535991.00)","0.42 (min 0.10, max 1.23)","1535992.20 (min 1535988.00, max 1535997.00)","0.12 (min 0.10, max 0.19)","1536000.67 (min 1535996.00, max 1536003.00)","0.12 (min 0.11, max 0.12)" + "88200","2822372.20 (min 2822363.00, max 2822378.00)","0.50 (min 0.45, max 0.56)","2822380.80 (min 2822374.00, max 2822389.00)","0.43 (min 0.40, max 0.46)","2822396.67 (min 2822387.00, max 2822403.00)","0.43 (min 0.41, max 0.45)" + "96000","3071956.20 (min 3071939.00, max 3071969.00)","0.37 (min 0.16, max 1.00)","3071971.00 (min 3071963.00, max 3071983.00)","0.35 (min 0.14, max 0.99)","3071978.00 (min 3071960.00, max 3071992.00)","0.18 (min 0.15, max 0.21)" .. csv-table:: Audio Playback :header: "Sampling Rate (Hz)","am62xx_lp_sk-fs: Throughput (bits/sec)","am62xx_lp_sk-fs: CPU Load (%)","am62xx_sk-fs: Throughput (bits/sec)","am62xx_sk-fs: CPU Load (%)","am62xxsip_sk-fs: Throughput (bits/sec)","am62xxsip_sk-fs: CPU Load (%)" - "11025","352945.43 (min 352945.00, max 352947.00)","0.13 (min 0.12, max 0.14)","352946.86 (min 352946.00, max 352949.00)","0.13 (min 0.09, max 0.17)","352946.13 (min 352945.00, max 352948.00)","0.19 (min 0.09, max 0.32)" - "16000","512212.00 (min 512211.00, max 512213.00)","0.20 (min 0.11, max 0.54)","512213.29 (min 512211.00, max 512216.00)","0.13 (min 0.08, max 0.30)","512213.25 (min 512211.00, max 512216.00)","0.20 (min 0.10, max 0.31)" - "22050","705887.71 (min 705885.00, max 705892.00)","0.16 (min 0.15, max 0.17)","705888.71 (min 705885.00, max 705891.00)","0.14 (min 0.11, max 0.21)","705888.88 (min 705888.00, max 705890.00)","0.21 (min 0.12, max 0.35)" - "24000","705889.86 (min 705888.00, max 705892.00)","0.18 (min 0.17, max 0.19)","705891.57 (min 705890.00, max 705893.00)","0.17 (min 0.12, max 0.22)","705891.63 (min 705889.00, max 705894.00)","0.24 (min 0.14, max 0.37)" - "32000","1024421.71 (min 1024419.00, max 1024423.00)","0.14 (min 0.13, max 0.15)","1024424.57 (min 1024422.00, max 1024426.00)","0.13 (min 0.09, max 0.21)","1024424.25 (min 1024420.00, max 1024429.00)","0.19 (min 0.10, max 0.30)" - "44100","1411749.57 (min 1411554.00, max 1411783.00)","0.23 (min 0.21, max 0.25)","1411755.43 (min 1411579.00, max 1411787.00)","0.20 (min 0.16, max 0.26)","1411785.25 (min 1411779.00, max 1411793.00)","0.27 (min 0.18, max 0.38)" - "48000","1536632.83 (min 1536629.00, max 1536635.00)","0.26 (min 0.11, max 0.69)","1536637.57 (min 1536633.00, max 1536640.00)","0.16 (min 0.09, max 0.21)","1536636.00 (min 1536630.00, max 1536640.00)","0.29 (min 0.11, max 0.45)" - "88200","2823558.00 (min 2823550.00, max 2823562.00)","0.39 (min 0.38, max 0.42)","2823567.00 (min 2823558.00, max 2823572.00)","0.37 (min 0.32, max 0.41)","2823443.00 (min 2822703.00, max 2823573.00)","0.44 (min 0.34, max 0.58)" - "96000","3073257.40 (min 3073251.00, max 3073262.00)","0.25 (min 0.20, max 0.28)","3073268.57 (min 3073258.00, max 3073274.00)","0.22 (min 0.17, max 0.28)","3073266.57 (min 3073255.00, max 3073276.00)","0.35 (min 0.20, max 0.63)" + "11025","352945.00 (min 352944.00, max 352946.00)","0.21 (min 0.15, max 0.29)","352946.60 (min 352945.00, max 352948.00)","0.19 (min 0.14, max 0.26)","352947.50 (min 352947.00, max 352948.00)","0.25 (min 0.21, max 0.28)" + "16000","512211.67 (min 512210.00, max 512213.00)","0.16 (min 0.10, max 0.32)","512213.20 (min 512212.00, max 512214.00)","0.32 (min 0.12, max 0.90)","512215.00 (min 512214.00, max 512216.00)","0.17 (min 0.14, max 0.19)" + "22050","705888.00 (min 705886.00, max 705890.00)","0.15 (min 0.13, max 0.17)","705890.20 (min 705888.00, max 705892.00)","0.13 (min 0.11, max 0.15)","705892.00 (min 705888.00, max 705896.00)","0.14 (min 0.13, max 0.14)" + "24000","705881.00 (min 705832.00, max 705891.00)","0.16 (min 0.13, max 0.18)","705881.80 (min 705840.00, max 705893.00)","0.14 (min 0.13, max 0.16)","705895.00 (min 705894.00, max 705896.00)","0.14 (min 0.13, max 0.15)" + "32000","1024422.50 (min 1024420.00, max 1024423.00)","0.12 (min 0.10, max 0.17)","1024425.20 (min 1024422.00, max 1024427.00)","0.12 (min 0.09, max 0.15)","1024429.50 (min 1024429.00, max 1024430.00)","0.12 (min 0.10, max 0.13)" + "44100","1411782.00 (min 1411778.00, max 1411783.00)","0.21 (min 0.19, max 0.22)","1411786.00 (min 1411781.00, max 1411788.00)","0.18 (min 0.16, max 0.21)","1411791.50 (min 1411790.00, max 1411793.00)","0.21 (min 0.20, max 0.22)" + "48000","1536632.80 (min 1536629.00, max 1536634.00)","0.13 (min 0.11, max 0.17)","1536637.75 (min 1536633.00, max 1536640.00)","0.11 (min 0.10, max 0.11)","1536644.50 (min 1536643.00, max 1536646.00)","0.63 (min 0.47, max 0.79)" + "88200","2823553.20 (min 2823535.00, max 2823561.00)","0.32 (min 0.29, max 0.35)","2823566.75 (min 2823558.00, max 2823572.00)","0.29 (min 0.25, max 0.33)","2823576.50 (min 2823575.00, max 2823578.00)","0.31 (min 0.29, max 0.33)" + "96000","3073248.50 (min 3073237.00, max 3073262.00)","0.29 (min 0.17, max 0.59)","3073254.67 (min 3073245.00, max 3073270.00)","0.17 (min 0.15, max 0.19)","3073253.00 (min 3073233.00, max 3073273.00)","0.39 (min 0.24, max 0.54)" | -Graphics SGX/RGX Driver ------------------------ +Graphics SGX/RGX +================ GFXBench -^^^^^^^^ +-------- + Run GFXBench and capture performance reported (Score and Display rate in fps). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: GFXBench Performance :header: "Benchmark","am62xx_lp_sk-fs: Score","am62xx_lp_sk-fs: Fps","am62xx_sk-fs: Score","am62xx_sk-fs: Fps" - " GFXBench 3.x gl_manhattan_off","81.66 (min 80.20, max 82.31)","1.32 (min 1.29, max 1.33)" - " GFXBench 3.x gl_trex_off","123.00 (min 120.86, max 125.32)","2.20 (min 2.16, max 2.24)" - " GFXBench 5.x gl_5_high_off","11.20 (min 11.14, max 11.28)","0.17 (min 0.17, max 0.18)","11.76 (min 11.60, max 11.89)","0.18" + " GFXBench 3.x gl_manhattan_off","81.53 (min 81.16, max 82.23)","1.32 (min 1.31, max 1.33)" + " GFXBench 3.x gl_trex_off","124.19 (min 122.95, max 124.90)","2.22 (min 2.20, max 2.23)" + " GFXBench 5.x gl_5_high_off","11.08 (min 10.87, max 11.19)","0.17","11.79 (min 11.60, max 11.89)","0.18" Glmark2 -^^^^^^^ +------- Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: Glmark2 Performance :header: "Benchmark","am62xx_lp_sk-fs: Score","am62xx_sk-fs: Score","am62xxsip_sk-fs: Score" - "Glmark2-DRM","59.00 (min 51.00, max 70.00)","61.00","54.00" - "Glmark2-Wayland","199.00 (min 198.00, max 203.00)","216.00","208.75 (min 208.00, max 209.00)" - "Glmark2-Wayland 1920x1080","62.57 (min 62.00, max 63.00)","66.29 (min 66.00, max 67.00)" + "Glmark2-DRM","57.00 (min 51.00, max 69.00)","61.00","54.00" + "Glmark2-Wayland","200.33 (min 199.00, max 203.00)","216.50 (min 216.00, max 217.00)","209.00" + "Glmark2-Wayland 1920x1080","62.67 (min 62.00, max 63.00)","67.00" | - Ethernet ------------------ -Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html +======== + +Ethernet performance benchmarks were measured using :command:`netperf` 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, @@ -453,11 +469,11 @@ CPU utilization was captured as the total percentage used across all cores on th while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. -In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth +In this scenario, :command:`netperf` options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: -:: +.. code-block:: console burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) burst_size = 500000000 / 8 / 1472 / 100 = 425 @@ -465,21 +481,21 @@ no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B data wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the :command:`netperf` test with no bandwidth limit (remove -b/-w options). -In order to start a netperf client on one device, the other device must have netserver running. -To start netserver: +In order to start a :command:`netperf` client on one device, the other device must have :command:`netserver` running. +To start :command:`netserver`: -:: +.. code-block:: console netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger :command:`netperf` clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test. -:: +.. code-block:: console #!/bin/bash for i in 1 @@ -491,102 +507,102 @@ additional timing measurements during the test. -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for -60 seconds at various burst/datagram sizes and report CPU utilization. +Running the following commands will trigger :command:`netperf` clients to measure UDP burst performance for +60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run :command:`netperf` client from DUT and start :command:`netserver` on tester. -:: +.. code-block:: console netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run :command:`netperf` client from tester and start :command:`netserver` on DUT. -:: +.. code-block:: console netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE +CPSW/CPSW2g/CPSW3g Ethernet +--------------------------- -CPSW/CPSW2g/CPSW3g Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -- CPSW2g: AM65x, J7200, J721e, J721S2, J784S4, J742S2 -- CPSW3g: AM64x, AM62x, AM62ax, AM62px +- CPSW3g: AM62x -.. rubric:: TCP Bidirectional Throughput - :name: CPSW2g-tcp-bidirectional-throughput +TCP Bidirectional Throughput +^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1698.44 (min 1628.06, max 1800.35)","67.30 (min 64.15, max 72.01)","1783.17 (min 1762.43, max 1798.15)","67.94 (min 65.84, max 69.35)","1658.66 (min 1180.00, max 1849.79)","66.19 (min 55.35, max 71.09)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1606.03 (min 1303.43, max 1816.77)","60.01 (min 39.33, max 71.84)","1540.48 (min 1283.30, max 1773.53)","52.49 (min 40.80, max 68.45)","1673.91 (min 1656.34, max 1691.48)","64.30 (min 62.64, max 65.95)" -.. rubric:: TCP Bidirectional Throughput Interrupt Pacing - :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing +TCP Bidirectional Throughput Interrupt Pacing +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1630.25 (min 1484.82, max 1709.62)","59.65 (min 50.43, max 84.96)","1604.35 (min 301.65, max 1837.36)","47.99 (min 14.99, max 55.09)","1840.19 (min 1813.97, max 1870.78)","54.44 (min 50.41, max 57.84)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1641.49 (min 1549.80, max 1756.23)","51.62 (min 39.44, max 63.10)","1563.91 (min 1444.87, max 1707.72)","39.79 (min 33.68, max 47.59)","1731.44 (min 1694.58, max 1768.29)","50.23 (min 42.61, max 57.85)" -.. rubric:: UDP Throughput - :name: CPSW2g-udp-throughput-0-loss +UDP Throughput +^^^^^^^^^^^^^^ .. csv-table:: CPSW2g UDP Egress Throughput 0 loss :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","45.17 (min 42.87, max 45.99)","88.22 (min 84.00, max 90.00)","39.84 (min 25.15, max 63.81)","","42.93 (min 26.73, max 54.34)","83.86 (min 52.00, max 106.00)","30.70 (min 12.22, max 39.15)","","47.46 (min 45.64, max 48.58)","92.75 (min 89.00, max 95.00)","39.02 (min 38.09, max 39.68)" - "128","82.00","87.45 (min 81.17, max 89.93)","85.25 (min 79.00, max 88.00)","44.32 (min 37.82, max 62.92)","","90.95 (min 17.31, max 102.75)","88.70 (min 17.00, max 100.00)","34.13 (min 9.70, max 38.73)","","95.20 (min 89.43, max 97.29)","93.00 (min 87.00, max 95.00)","39.06 (min 38.06, max 39.81)" - "256","210.00","172.10 (min 156.28, max 178.99)","83.89 (min 76.00, max 87.00)","43.40 (min 37.59, max 62.50)","","200.58 (min 195.49, max 213.41)","97.71 (min 95.00, max 104.00)","38.57 (min 38.11, max 39.49)","","185.76 (min 178.48, max 190.50)","90.57 (min 87.00, max 93.00)","38.56 (min 38.29, max 38.82)" - "1024","978.00","610.03 (min 286.71, max 692.59)","74.57 (min 35.00, max 85.00)","37.81 (min 15.50, max 61.34)","","627.48 (min 46.69, max 807.16)","76.78 (min 6.00, max 99.00)","30.94 (min 0.92, max 38.45)","","700.39 (min 683.61, max 736.41)","85.25 (min 83.00, max 90.00)","38.01 (min 37.61, max 38.45)" - "1518","1472.00","530.95 (min 183.19, max 697.98)","43.63 (min 15.00, max 57.00)","35.51 (min 11.05, max 60.04)","","757.07 (min 611.07, max 823.49)","62.33 (min 50.00, max 68.00)","35.49 (min 32.74, max 36.36)","","692.69 (min 529.45, max 741.46)","57.13 (min 44.00, max 61.00)","35.56 (min 30.07, max 40.87)" + "64","","46.06 (min 37.80, max 54.51)","90.00 (min 74.00, max 106.00)","39.08 (min 36.43, max 42.06)","","53.38 (min 42.81, max 62.88)","104.40 (min 84.00, max 123.00)","39.68 (min 36.73, max 42.59)","","44.24 (min 43.33, max 45.15)","86.50 (min 85.00, max 88.00)","37.12 (min 36.93, max 37.30)" + "128","","85.61 (min 73.61, max 109.51)","83.67 (min 72.00, max 107.00)","38.00 (min 36.03, max 41.81)","","104.87 (min 82.30, max 121.26)","102.20 (min 80.00, max 118.00)","39.45 (min 36.24, max 41.89)","","90.28 (min 89.44, max 91.11)","88.00 (min 87.00, max 89.00)","37.38 (min 37.37, max 37.38)" + "256","","179.15 (min 147.10, max 213.03)","87.50 (min 72.00, max 104.00)","38.96 (min 36.23, max 42.27)","","127.26 (min 53.04, max 239.91)","62.00 (min 26.00, max 117.00)","24.61 (min 8.86, max 41.07)","","174.19 (min 171.67, max 176.70)","85.00 (min 84.00, max 86.00)","36.85 (min 36.54, max 37.16)" + "1024","","662.53 (min 586.81, max 812.44)","81.00 (min 72.00, max 99.00)","37.83 (min 35.95, max 40.50)","","395.91 (min 163.01, max 663.99)","48.40 (min 20.00, max 81.00)","20.92 (min 8.11, max 36.34)","","645.78 (min 634.74, max 656.81)","78.50 (min 77.00, max 80.00)","36.08 (min 35.77, max 36.38)" + "1518","","700.49 (min 612.17, max 810.09)","57.75 (min 50.00, max 67.00)","36.31 (min 34.89, max 37.94)","","606.38 (min 537.43, max 675.32)","50.00 (min 44.00, max 56.00)","32.91 (min 30.89, max 34.93)","","699.12 (min 692.68, max 705.56)","57.50 (min 57.00, max 58.00)","35.18 (min 35.07, max 35.28)" .. csv-table:: CPSW2g UDP Ingress Throughput 0 loss :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","2.27 (min 1.38, max 2.66)","4.33 (min 3.00, max 5.00)","1.77 (min 0.74, max 3.49)","","1.52 (min 1.38, max 1.54)","3.00","0.88 (min 0.62, max 1.82)","","1.92 (min 1.38, max 2.71)","3.89 (min 3.00, max 5.00)","1.18 (min 0.58, max 2.47)" - "128","82.00","5.28 (min 5.12, max 5.43)","5.00","2.32 (min 1.08, max 4.15)","","4.51","4.00","1.19 (min 0.85, max 2.88)","","4.73 (min 4.10, max 5.32)","4.57 (min 4.00, max 5.00)","0.99 (min 0.76, max 1.18)" - "256","210.00","10.38 (min 9.83, max 10.85)","5.00","2.01 (min 1.04, max 3.87)","","11.06 (min 9.83, max 16.18)","5.50 (min 5.00, max 8.00)","1.76 (min 1.05, max 3.74)","","10.32 (min 9.20, max 10.85)","4.80 (min 4.00, max 5.00)","1.06 (min 0.98, max 1.15)" - "1024","978.00","43.15 (min 42.60, max 43.42)","5.00","3.05 (min 1.38, max 5.13)","","43.07 (min 42.60, max 44.23)","5.00","2.44 (min 1.48, max 3.70)","","189.01 (min 27.85, max 935.80)","22.83 (min 3.00, max 114.00)","8.78 (min 1.39, max 42.77)" - "1518","1472.00","61.97 (min 61.23, max 62.41)","5.00","3.53 (min 1.95, max 4.71)","","62.18 (min 61.23, max 63.59)","5.00","3.07 (min 1.84, max 4.74)","","225.77 (min 28.26, max 941.32)","19.00 (min 2.00, max 80.00)","11.77 (min 1.97, max 42.15)" + "64","","1.82 (min 1.38, max 2.10)","3.75 (min 3.00, max 4.00)","2.15 (min 1.69, max 2.84)","","1.72 (min 1.43, max 2.10)","3.40 (min 3.00, max 4.00)","1.42 (min 0.53, max 2.65)","","2.18 (min 2.10, max 2.25)","4.00","1.24 (min 1.04, max 1.43)" + "128","","4.07 (min 3.69, max 5.02)","4.25 (min 4.00, max 5.00)","2.53 (min 1.33, max 4.89)","","4.55 (min 4.10, max 5.22)","4.40 (min 4.00, max 5.00)","2.22 (min 0.81, max 4.32)","","4.86 (min 4.40, max 5.32)","4.50 (min 4.00, max 5.00)","1.28 (min 0.97, max 1.58)" + "256","","9.73 (min 9.22, max 10.24)","5.00","3.26 (min 2.02, max 4.50)","","9.63 (min 9.01, max 10.24)","4.67 (min 4.00, max 5.00)","1.47 (min 1.21, max 1.89)","","10.04 (min 9.83, max 10.24)","5.00","1.66 (min 1.21, max 2.11)" + "1024","","41.78","5.00","2.92 (min 2.18, max 4.20)","","41.78","5.00","1.58 (min 1.40, max 1.90)","","41.78","5.00","2.82" + "1518","","61.23","5.00","3.88 (min 3.72, max 4.08)","","70.26 (min 60.06, max 89.50)","6.00 (min 5.00, max 8.00)","4.64 (min 4.27, max 5.26)" .. csv-table:: CPSW2g UDP Ingress Throughput possible loss :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_lp_sk-fs: Packet Loss %","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: Packet Loss %","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: Packet Loss %" - "64","18.00","78.62 (min 69.58, max 88.72)","153.56 (min 136.00, max 173.00)","48.04 (min 39.65, max 69.17)","52.00 (min 0.03, max 80.01)","","96.29 (min 89.66, max 97.42)","188.13 (min 175.00, max 190.00)","42.55 (min 39.69, max 44.43)","47.53 (min 31.90, max 64.49)","","83.77 (min 61.91, max 91.66)","163.56 (min 121.00, max 179.00)","42.08 (min 40.20, max 43.92)","71.27 (min 65.75, max 77.58)" - "128","82.00","166.41 (min 141.23, max 171.82)","162.57 (min 138.00, max 168.00)","43.50 (min 40.85, max 45.21)","45.31 (min 0.18, max 62.80)","","192.35 (min 182.52, max 194.35)","187.89 (min 178.00, max 190.00)","40.78 (min 33.34, max 46.88)","48.04 (min 28.69, max 70.54)","","170.66 (min 161.89, max 179.10)","166.71 (min 158.00, max 175.00)","43.01 (min 41.89, max 44.60)","66.48 (min 46.79, max 74.38)" - "256","210.00","302.93 (min 274.45, max 334.43)","147.89 (min 134.00, max 163.00)","48.96 (min 40.52, max 69.72)","28.02 (min 0.35, max 61.23)","","365.04 (min 343.99, max 374.14)","178.33 (min 168.00, max 183.00)","43.26 (min 41.74, max 44.53)","47.95 (min 40.06, max 53.04)","","285.42 (min 148.67, max 331.95)","139.40 (min 73.00, max 162.00)","42.81 (min 39.40, max 44.21)","46.03 (min 16.49, max 58.26)" - "1024","978.00","815.64 (min 461.33, max 935.09)","99.56 (min 56.00, max 114.00)","44.65 (min 38.28, max 68.37)","0.31 (min 0.07, max 0.81)","","889.38 (min 838.56, max 935.45)","108.43 (min 102.00, max 114.00)","42.69 (min 39.75, max 45.31)","0.43 (min 0.06, max 0.90)","","817.58 (min 732.19, max 908.97)","99.83 (min 89.00, max 111.00)","41.74 (min 40.22, max 43.02)","0.48 (min 0.09, max 0.89)" - "1518","1472.00","847.13 (min 748.02, max 945.15)","72.00 (min 64.00, max 80.00)","43.54 (min 35.25, max 68.38)","0.43 (min 0.03, max 1.40)","","881.34 (min 797.89, max 922.82)","74.80 (min 68.00, max 78.00)","40.02 (min 35.44, max 42.18)","0.34 (min 0.16, max 0.49)","","922.03 (min 879.70, max 955.84)","78.29 (min 75.00, max 81.00)","41.81 (min 39.60, max 43.08)","0.41 (min 0.00, max 0.77)" + "64","","69.11 (min 59.97, max 87.93)","135.00 (min 117.00, max 172.00)","36.72 (min 34.78, max 38.94)","26.42 (min 1.56, max 65.56)","","82.74 (min 67.76, max 97.96)","161.40 (min 132.00, max 191.00)","38.08 (min 34.78, max 42.57)","63.67 (min 43.77, max 74.38)","","69.59 (min 67.60, max 71.58)","136.00 (min 132.00, max 140.00)","34.70 (min 34.39, max 35.00)","74.32 (min 68.16, max 80.48)" + "128","","141.22 (min 119.76, max 162.02)","137.75 (min 117.00, max 158.00)","38.88 (min 36.11, max 42.02)","52.69 (min 20.71, max 75.48)","","158.34 (min 132.52, max 176.61)","154.60 (min 129.00, max 172.00)","37.12 (min 35.20, max 39.85)","62.73 (min 53.38, max 72.34)","","138.96 (min 136.23, max 141.68)","135.50 (min 133.00, max 138.00)","35.06 (min 34.90, max 35.22)","76.84 (min 76.16, max 77.52)" + "256","","258.83 (min 235.45, max 282.21)","126.50 (min 115.00, max 138.00)","37.83 (min 36.04, max 39.62)","11.20 (min 1.41, max 20.99)","","332.00 (min 263.81, max 384.23)","162.33 (min 129.00, max 188.00)","40.55 (min 35.77, max 46.01)","53.96 (min 50.78, max 58.00)","","255.67 (min 249.57, max 261.77)","125.00 (min 122.00, max 128.00)","35.55 (min 35.28, max 35.81)","57.33 (min 56.21, max 58.44)" + "1024","","776.19 (min 642.99, max 892.17)","94.67 (min 78.00, max 109.00)","38.97 (min 36.84, max 41.48)","3.63 (min 1.35, max 5.01)","","929.77 (min 919.63, max 936.99)","113.25 (min 112.00, max 114.00)","39.72 (min 37.19, max 42.42)","0.70 (min 0.25, max 1.35)","","676.55","83.00","34.96","1.77" + "1518","","875.44 (min 744.64, max 951.13)","74.33 (min 63.00, max 81.00)","40.26 (min 38.35, max 41.54)","0.78 (min 0.56, max 1.15)","","943.68 (min 942.65, max 945.04)","80.00","40.37 (min 39.53, max 40.82)","1.27 (min 0.89, max 1.50)" | -Linux OSPI Flash Driver ------------------------ +OSPI +==== -.. rubric:: AM62XX-SK - :name: am62xx-sk-ospi-flash-driver - -.. rubric:: RAW - :name: am62xx-sk-ospi-raw +OSPI RAW +-------- .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62xx_sk-fs: Raw Read Throughput (Mbytes/sec)" - "50","170.77 (min 166.67, max 172.41)" - -.. rubric:: RAW - :name: am62xx-sk-lp-ospi-raw + "50","151.67 (min 21.01, max 192.31)" .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62xx_lp_sk-fs: Raw Read Throughput (Mbytes/sec)" - "50","29.24 (min 28.73, max 29.59)" + "50","30.49 (min 29.94, max 31.05)" + +.. csv-table:: OSPI Raw Flash Driver + :header: "File size (Mbytes)","am62xxsip_sk-fs: Raw Read Throughput (Mbytes/sec)" + + "50","108.69" + +| + +EMMC +==== -EMMC Driver ------------ .. warning:: **IMPORTANT**: The performance numbers can be severely affected if the media is @@ -596,115 +612,116 @@ EMMC Driver re-mount in async mode. EMMC EXT4 FIO 1G -^^^^^^^^^^^^^^^^ +---------------- .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "1m","55.21 (min 43.80, max 90.90)","1.48 (min 1.08, max 2.52)","174.00 (min 171.00, max 175.00)","2.05 (min 1.79, max 2.27)" - "4m","56.05 (min 43.70, max 96.40)","1.11 (min 0.80, max 1.78)","174.00 (min 171.00, max 175.00)","1.65 (min 1.46, max 1.87)" - "4k","29.97 (min 5.31, max 63.60)","10.46 (min 2.06, max 22.72)","50.86 (min 36.30, max 93.30)","13.38 (min 9.65, max 23.05)" - "256k","50.74 (min 35.00, max 91.30)","1.69 (min 1.11, max 2.90)","173.63 (min 171.00, max 174.00)","2.68 (min 2.25, max 2.96)" + "1m","56.35 (min 44.40, max 75.60)","1.56 (min 1.16, max 2.04)","164.17 (min 111.00, max 175.00)","1.90 (min 1.59, max 2.12)" + "4m","56.83 (min 44.50, max 80.10)","1.06 (min 0.71, max 1.42)","164.00 (min 110.00, max 175.00)","1.54 (min 1.18, max 1.84)" + "4k","28.60 (min 5.36, max 52.00)","9.06 (min 1.93, max 15.76)","49.53 (min 36.40, max 56.10)","12.30 (min 9.51, max 14.25)" + "256k","52.10 (min 35.10, max 67.00)","1.66 (min 1.12, max 2.17)","160.28 (min 91.70, max 174.00)","2.61 (min 1.70, max 2.93)" .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "1m","49.95 (min 42.90, max 58.90)","1.58 (min 1.28, max 1.85)","174.50 (min 174.00, max 175.00)","2.37 (min 1.96, max 2.55)" - "4m","49.89 (min 43.10, max 58.90)","1.27 (min 0.98, max 1.86)","174.70 (min 174.00, max 175.00)","1.91 (min 1.63, max 2.35)" - "4k","23.80 (min 5.24, max 51.70)","9.27 (min 2.38, max 19.72)","44.31 (min 36.40, max 56.30)","13.52 (min 11.10, max 16.93)" - "256k","44.37 (min 34.20, max 58.80)","1.75 (min 1.31, max 2.31)","174.00","3.09 (min 2.66, max 3.23)" + "1m","55.60 (min 45.30, max 61.60)","1.78 (min 1.52, max 2.04)","174.75 (min 174.00, max 175.00)","2.37 (min 2.24, max 2.58)" + "4m","55.50 (min 45.40, max 61.20)","1.29 (min 1.18, max 1.45)","174.75 (min 174.00, max 175.00)","1.92 (min 1.81, max 2.02)" + "4k","39.94 (min 5.26, max 51.60)","14.36 (min 2.35, max 18.53)","51.20 (min 36.40, max 56.80)","14.69 (min 10.45, max 16.16)" + "256k","53.20 (min 35.80, max 61.50)","2.10 (min 1.60, max 2.34)","174.00","3.19 (min 2.86, max 3.41)" .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write EXT4 CPU Load (%)","am62xxsip_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read EXT4 CPU Load (%)" - "1m","92.76 (min 91.00, max 95.90)","2.53 (min 2.25, max 2.94)","171.75 (min 171.00, max 174.00)","2.55 (min 2.22, max 2.87)" - "4m","96.81 (min 92.20, max 98.00)","2.27 (min 2.05, max 2.53)","165.75 (min 141.00, max 173.00)","2.53 (min 2.11, max 2.84)" - "4k","64.16 (min 63.70, max 64.50)","24.79 (min 24.52, max 25.21)","92.38 (min 91.60, max 92.80)","24.06 (min 23.88, max 24.22)" - "256k","91.96 (min 91.00, max 92.80)","2.63 (min 2.38, max 2.93)","171.00 (min 170.00, max 173.00)","2.92 (min 2.64, max 3.14)" + "1m","91.63 (min 91.30, max 92.00)","2.51 (min 2.23, max 2.67)","171.33 (min 170.00, max 173.00)","2.39 (min 2.14, max 2.57)" + "4m","97.27 (min 96.90, max 97.70)","2.12 (min 1.89, max 2.31)","167.00 (min 158.00, max 172.00)","2.26 (min 1.93, max 2.48)" + "4k","63.67 (min 63.20, max 64.00)","20.73 (min 19.98, max 21.12)","93.40 (min 92.90, max 94.10)","22.07 (min 21.76, max 22.27)" + "256k","91.67 (min 91.00, max 92.50)","2.43 (min 2.08, max 2.63)","170.33 (min 170.00, max 171.00)","2.63 (min 2.35, max 2.79)" EMMC EXT4 -^^^^^^^^^ +--------- .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "102400","47.27 (min 39.38, max 68.59)","4.63 (min 3.08, max 8.55)","172.03 (min 142.58, max 179.51)","12.04 (min 10.43, max 14.23)" - "262144","46.33 (min 39.08, max 69.36)","4.62 (min 3.09, max 8.92)","171.71 (min 98.67, max 183.00)","12.59 (min 5.84, max 14.54)" - "524288","46.07 (min 39.39, max 69.29)","4.66 (min 3.27, max 8.86)","176.20 (min 132.59, max 182.99)","12.90 (min 9.65, max 14.54)" - "1048576","45.97 (min 38.94, max 67.84)","4.48 (min 3.09, max 8.61)","177.23 (min 139.09, max 182.88)","13.66 (min 10.74, max 15.35)" - "5242880","46.19 (min 39.09, max 68.52)","4.48 (min 2.89, max 8.63)","176.56 (min 136.16, max 182.81)","13.55 (min 11.04, max 15.65)" + "102400","50.31 (min 41.84, max 56.05)","3.33 (min 2.60, max 4.79)","180.62 (min 180.33, max 181.00)","8.68 (min 7.83, max 9.52)" + "262144","51.08 (min 42.48, max 56.58)","2.80 (min 2.15, max 4.33)","182.99 (min 182.78, max 183.18)","9.97 (min 8.41, max 10.96)" + "524288","51.06 (min 41.99, max 56.65)","2.73 (min 2.06, max 4.58)","182.90 (min 182.74, max 183.04)","10.36 (min 8.77, max 11.74)" + "1048576","50.98 (min 42.13, max 56.54)","2.65 (min 1.94, max 3.79)","182.93 (min 182.82, max 183.13)","9.78 (min 7.93, max 10.53)" + "5242880","51.07 (min 42.13, max 56.54)","2.75 (min 2.01, max 4.09)","182.78 (min 182.36, max 183.06)","10.78 (min 8.77, max 11.40)" .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "102400","46.20 (min 39.42, max 53.63)","5.41 (min 3.80, max 7.72)","175.55 (min 167.93, max 177.22)","14.04 (min 13.30, max 14.77)" - "262144","45.84 (min 39.01, max 52.84)","5.56 (min 4.45, max 7.78)","180.66 (min 180.04, max 181.28)","16.55 (min 15.93, max 17.83)" - "524288","45.78 (min 39.45, max 52.65)","5.57 (min 4.50, max 7.95)","182.34 (min 181.72, max 182.77)","17.30 (min 16.67, max 18.26)" - "1048576","45.98 (min 39.10, max 53.29)","5.48 (min 4.24, max 7.77)","181.96 (min 181.42, max 182.45)","17.45 (min 16.74, max 19.48)" - "5242880","46.10 (min 39.39, max 52.97)","5.41 (min 4.42, max 7.81)","181.51 (min 180.56, max 182.12)","17.27 (min 16.37, max 18.26)" + "102400","51.61 (min 43.30, max 56.17)","3.71 (min 2.43, max 5.02)","179.02 (min 172.26, max 180.85)","9.05 (min 8.23, max 9.57)" + "262144","52.59 (min 43.98, max 57.20)","3.15 (min 1.80, max 4.78)","183.01 (min 182.93, max 183.06)","10.99 (min 9.65, max 12.33)" + "524288","52.56 (min 43.72, max 57.24)","3.14 (min 2.01, max 4.45)","182.84 (min 182.51, max 183.01)","11.49 (min 10.04, max 13.54)" + "1048576","52.58 (min 43.57, max 57.28)","3.06 (min 1.90, max 4.24)","182.79 (min 182.41, max 182.96)","11.33 (min 10.48, max 12.72)" + "5242880","52.15 (min 43.78, max 55.83)","3.15 (min 2.01, max 4.62)","182.58 (min 181.98, max 182.86)","12.35 (min 10.87, max 13.16)" .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write EXT4 CPU Load (%)","am62xxsip_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read EXT4 CPU Load (%)" - "102400","92.90 (min 35.38, max 100.03)","12.70 (min 5.91, max 26.68)","178.51 (min 178.00, max 178.87)","12.36 (min 10.82, max 14.41)" - "262144","78.13 (min 30.00, max 97.55)","7.26 (min 2.26, max 17.43)","180.57 (min 180.08, max 181.05)","13.93 (min 12.66, max 15.42)" - "524288","77.25 (min 32.87, max 96.72)","6.82 (min 2.00, max 12.76)","181.08 (min 179.75, max 181.56)","14.21 (min 12.61, max 16.81)" - "1048576","75.27 (min 32.45, max 97.43)","6.62 (min 2.05, max 12.64)","181.08 (min 180.11, max 181.60)","13.54 (min 12.28, max 16.67)" - "5242880","73.92 (min 28.97, max 96.92)","6.62 (min 2.03, max 13.75)","180.86 (min 179.21, max 181.48)","13.65 (min 12.55, max 15.95)" + "102400","95.17 (min 78.78, max 99.27)","10.85 (min 3.48, max 20.87)","162.54 (min 159.41, max 166.83)","23.53 (min 19.52, max 30.28)" + "262144","74.93 (min 29.32, max 99.56)","5.48 (min 1.62, max 13.98)","174.98 (min 171.54, max 176.93)","22.29 (min 17.17, max 29.67)" + "524288","75.89 (min 30.26, max 99.36)","7.44 (min 2.98, max 18.12)","178.63 (min 177.90, max 180.04)","22.39 (min 17.60, max 29.24)" + "1048576","75.87 (min 33.26, max 97.92)","11.60 (min 1.61, max 23.10)","178.97 (min 178.01, max 180.14)","21.34 (min 18.70, max 24.68)" + "5242880","73.34 (min 11.84, max 97.06)","13.62 (min 3.25, max 24.94)","179.63 (min 178.91, max 180.27)","22.32 (min 19.21, max 26.81)" EMMC VFAT -^^^^^^^^^ +--------- .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62xx_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62xx_sk-fs: Write VFAT CPU Load (%)","am62xx_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62xx_sk-fs: Read VFAT CPU Load (%)" - "102400","41.99 (min 33.56, max 50.47)","5.63 (min 3.71, max 8.96)","168.88 (min 167.46, max 171.01)","12.46 (min 11.43, max 13.47)" - "262144","43.21 (min 34.82, max 51.89)","5.89 (min 4.06, max 8.84)","170.74 (min 170.03, max 172.53)","14.81 (min 12.50, max 15.98)" - "524288","43.26 (min 34.68, max 51.97)","5.92 (min 3.98, max 9.32)","169.14 (min 168.50, max 170.91)","14.50 (min 12.65, max 15.66)" - "1048576","43.40 (min 35.02, max 52.21)","5.83 (min 3.98, max 9.22)","168.62 (min 167.68, max 169.36)","14.54 (min 12.60, max 15.73)" - "5242880","43.37 (min 34.99, max 52.13)","5.80 (min 3.91, max 9.17)","168.87 (min 168.10, max 170.34)","14.71 (min 11.93, max 16.06)" + "102400","52.56 (min 34.40, max 71.36)","6.92 (min 3.55, max 12.09)","162.41 (min 152.54, max 173.22)","11.50 (min 9.89, max 12.79)" + "262144","55.10 (min 35.68, max 77.62)","7.45 (min 3.83, max 13.09)","163.45 (min 156.32, max 173.46)","13.60 (min 11.54, max 16.09)" + "524288","54.75 (min 35.78, max 77.29)","7.34 (min 4.04, max 12.59)","165.21 (min 151.84, max 173.93)","13.71 (min 12.30, max 15.42)" + "1048576","54.97 (min 35.68, max 78.17)","7.30 (min 3.98, max 12.78)","167.79 (min 155.51, max 173.76)","14.55 (min 13.36, max 15.83)" + "5242880","54.86 (min 35.78, max 77.98)","7.34 (min 3.89, max 12.59)","165.91 (min 153.32, max 173.53)","14.65 (min 13.36, max 16.05)" .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write VFAT CPU Load (%)","am62xx_lp_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read VFAT CPU Load (%)" - "102400","40.98 (min 31.93, max 49.66)","6.24 (min 4.49, max 9.95)","167.36 (min 165.22, max 169.10)","14.86 (min 13.22, max 16.27)" - "262144","42.56 (min 33.32, max 51.62)","6.69 (min 4.94, max 10.29)","169.22 (min 167.97, max 170.23)","17.21 (min 15.42, max 18.22)" - "524288","42.57 (min 33.42, max 51.68)","6.73 (min 5.00, max 10.84)","167.98 (min 167.06, max 168.70)","16.59 (min 15.51, max 17.67)" - "1048576","42.66 (min 33.48, max 51.79)","6.55 (min 4.89, max 9.98)","167.73 (min 167.03, max 168.46)","16.94 (min 15.85, max 17.74)" - "5242880","42.78 (min 33.53, max 52.13)","6.57 (min 4.94, max 10.26)","167.31 (min 166.74, max 168.09)","16.85 (min 15.73, max 17.93)" + "102400","43.14 (min 31.28, max 48.03)","6.67 (min 4.77, max 9.79)","168.45 (min 168.10, max 168.97)","14.66 (min 13.88, max 15.38)" + "262144","44.58 (min 32.66, max 49.74)","6.98 (min 4.81, max 10.22)","169.28 (min 167.93, max 170.55)","17.06 (min 16.26, max 17.55)" + "524288","44.58 (min 32.73, max 49.63)","7.11 (min 4.99, max 12.17)","167.27 (min 166.39, max 168.76)","16.65 (min 15.73, max 17.67)" + "1048576","44.59 (min 31.62, max 50.06)","6.88 (min 4.70, max 10.08)","166.64 (min 165.59, max 167.41)","15.93 (min 14.46, max 17.53)" + "5242880","43.91 (min 31.57, max 50.45)","6.84 (min 4.60, max 10.32)","165.31 (min 164.67, max 166.08)","17.23 (min 16.21, max 17.79)" .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write VFAT CPU Load (%)","am62xxsip_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read VFAT CPU Load (%)" - "102400","80.24 (min 45.77, max 96.90)","16.11 (min 5.36, max 29.32)","174.50 (min 173.84, max 174.86)","12.42 (min 10.59, max 15.29)" - "262144","73.99 (min 51.85, max 96.56)","7.98 (min 4.06, max 15.38)","175.43 (min 175.16, max 175.60)","13.91 (min 12.71, max 17.08)" - "524288","72.90 (min 49.94, max 96.87)","7.76 (min 3.53, max 14.31)","174.89 (min 174.56, max 175.13)","13.82 (min 12.97, max 15.90)" - "1048576","74.90 (min 49.90, max 96.59)","8.42 (min 4.81, max 15.75)","174.62 (min 174.19, max 174.99)","13.74 (min 12.66, max 16.60)" - "5242880","75.79 (min 49.91, max 96.31)","8.88 (min 4.68, max 13.65)","174.76 (min 174.11, max 175.19)","13.69 (min 12.61, max 15.61)" + "102400","83.18 (min 56.50, max 98.42)","12.41 (min 7.21, max 20.04)","174.58 (min 173.60, max 175.16)","10.94 (min 10.92, max 10.97)" + "262144","76.86 (min 58.60, max 98.13)","7.57 (min 4.32, max 13.34)","175.43 (min 174.85, max 175.83)","12.22 (min 11.76, max 12.71)" + "524288","71.89 (min 52.63, max 97.46)","6.91 (min 3.72, max 12.58)","175.03 (min 174.75, max 175.37)","12.41 (min 11.76, max 13.33)" + "1048576","79.09 (min 59.19, max 96.71)","7.86 (min 3.77, max 12.68)","174.88 (min 174.20, max 175.33)","13.16 (min 11.76, max 15.48)" + "5242880","78.00 (min 58.68, max 98.05)","8.31 (min 4.45, max 13.28)","174.56 (min 174.33, max 174.70)","13.21 (min 12.97, max 13.39)" + +UBoot EMMC +---------- -UBoot EMMC Driver ------------------ .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62xx_sk-fs: Write Throughput (Kbytes/sec)","am62xx_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","53539.31 (min 13727.69, max 61134.33)","170336.22 (min 169782.38, max 171560.21)" - "4000000","64033.22 (min 58514.29, max 95533.53)","172287.23 (min 164663.32, max 173835.54)" + "2000000","64931.38 (min 56205.83, max 98107.78)","169639.54 (min 168041.03, max 170666.67)" + "4000000","65572.89 (min 58776.68, max 96376.47)","172398.70 (min 169343.67, max 173375.66)" .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62xx_lp_sk-fs: Write Throughput (Kbytes/sec)","am62xx_lp_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","60091.96 (min 57186.74, max 62178.37)","170371.91 (min 169782.38, max 170666.67)" - "4000000","61147.19 (min 57893.99, max 62415.24)","165622.92 (min 100824.62, max 174297.87)" + "2000000","59176.51 (min 56399.31, max 61593.98)","170845.38 (min 170666.67, max 171560.21)" + "4000000","60389.18 (min 58202.49, max 62178.37)","174297.87" .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62xxsip_sk-fs: Write Throughput (Kbytes/sec)","am62xxsip_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","86450.84 (min 13914.23, max 97523.81)","163230.55 (min 141852.81, max 172463.16)" - "4000000","98727.53 (min 96234.95, max 101448.92)","173236.44 (min 158682.81, max 175229.95)" + "2000000","97208.26 (min 93090.91, max 102721.00)","171861.19 (min 171560.21, max 172463.16)" + "4000000","98611.89 (min 97234.42, max 99902.44)","175229.95" MMCSD ------ +===== .. warning:: @@ -715,61 +732,61 @@ MMCSD re-mount in async mode. MMC EXT4 FIO 1G -^^^^^^^^^^^^^^^ +--------------- .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "1m","42.31 (min 40.80, max 43.60)","1.53 (min 1.24, max 1.69)","87.53 (min 87.30, max 88.20)","1.59 (min 1.34, max 1.72)" - "4m","41.66 (min 40.40, max 42.40)","1.03 (min 0.92, max 1.11)","86.67 (min 82.60, max 87.50)","1.19 (min 0.94, max 1.27)" - "4k","2.79 (min 2.76, max 2.83)","1.88 (min 1.71, max 1.99)","12.93 (min 12.80, max 13.00)","4.47 (min 4.27, max 4.71)" - "256k","38.07 (min 36.30, max 39.10)","1.70 (min 1.40, max 1.85)","83.87 (min 83.60, max 84.30)","1.75 (min 1.60, max 1.87)" + "1m","42.46 (min 41.90, max 43.50)","1.47 (min 1.10, max 1.67)","87.54 (min 87.30, max 88.10)","1.59 (min 1.47, max 1.75)" + "4m","41.78 (min 40.20, max 43.20)","0.99 (min 0.81, max 1.13)","86.26 (min 82.50, max 87.30)","1.13 (min 1.04, max 1.24)" + "4k","2.82 (min 2.78, max 2.87)","1.76 (min 1.63, max 1.88)","12.92 (min 12.80, max 13.00)","4.41 (min 4.30, max 4.51)" + "256k","38.54 (min 37.40, max 40.00)","1.62 (min 1.20, max 1.79)","83.90 (min 83.50, max 84.60)","1.77 (min 1.66, max 1.90)" .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "1m","42.51 (min 41.90, max 43.60)","1.66 (min 1.59, max 1.74)","87.04 (min 84.90, max 87.40)","1.74 (min 1.66, max 1.87)" - "4m","42.10 (min 40.70, max 43.20)","1.16 (min 1.07, max 1.25)","86.80 (min 84.20, max 87.40)","1.27 (min 1.22, max 1.31)" - "4k","2.79 (min 2.76, max 2.84)","2.12 (min 1.97, max 2.26)","12.90 (min 12.80, max 13.00)","4.96 (min 4.77, max 5.13)" - "256k","37.40 (min 36.10, max 38.50)","1.83 (min 1.64, max 1.93)","83.70 (min 83.30, max 84.30)","1.99 (min 1.87, max 2.06)" + "1m","43.30 (min 42.60, max 43.60)","1.68 (min 1.55, max 1.78)","87.23 (min 85.30, max 88.10)","1.73 (min 1.65, max 1.84)" + "4m","41.55 (min 40.80, max 43.00)","1.10 (min 1.05, max 1.16)","84.75 (min 82.40, max 87.40)","1.31 (min 1.27, max 1.37)" + "4k","2.85 (min 2.84, max 2.86)","1.97 (min 1.88, max 2.03)","13.03 (min 12.90, max 13.10)","4.71 (min 4.63, max 4.77)" + "256k","38.03 (min 35.90, max 39.20)","1.76 (min 1.63, max 1.88)","84.18 (min 83.70, max 84.50)","2.03 (min 2.02, max 2.04)" .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write EXT4 CPU Load (%)","am62xxsip_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read EXT4 CPU Load (%)" - "1m","39.39 (min 18.80, max 42.80)","1.61 (min 1.16, max 1.88)","87.29 (min 86.60, max 87.70)","1.91 (min 1.67, max 2.13)" - "4m","38.44 (min 18.60, max 42.50)","1.34 (min 0.85, max 1.57)","87.00 (min 86.20, max 87.30)","1.78 (min 1.37, max 2.02)" - "4k","2.96 (min 2.77, max 4.18)","2.17 (min 2.04, max 2.73)","13.00 (min 12.90, max 13.50)","4.93 (min 4.70, max 5.20)" - "256k","35.65 (min 17.70, max 39.60)","1.69 (min 1.15, max 1.92)","83.46 (min 82.20, max 84.00)","1.91 (min 1.68, max 2.08)" + "1m","35.27 (min 19.60, max 43.90)","1.37 (min 1.08, max 1.62)","87.33 (min 87.20, max 87.50)","1.56 (min 1.53, max 1.60)" + "4m","34.37 (min 19.10, max 42.90)","1.01 (min 0.73, max 1.16)","87.27 (min 87.20, max 87.40)","1.51 (min 1.38, max 1.75)" + "4k","3.24 (min 2.79, max 4.10)","1.86 (min 1.57, max 2.38)","13.13 (min 12.90, max 13.60)","4.24 (min 4.10, max 4.48)" + "256k","32.00 (min 18.90, max 39.70)","1.36 (min 1.21, max 1.47)","83.60 (min 83.20, max 83.90)","1.58 (min 1.40, max 1.80)" MMC EXT4 -^^^^^^^^ +-------- .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62xx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Write Raw CPU Load (%)","am62xx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Read Raw CPU Load (%)" - "102400","10.57 (min 10.18, max 10.82)","1.12 (min 0.89, max 1.52)","11.02 (min 10.86, max 11.09)","0.96 (min 0.85, max 1.08)" - "262144","10.35 (min 10.13, max 10.84)","1.14 (min 0.84, max 1.65)","11.11 (min 11.00, max 11.22)","1.02 (min 0.80, max 1.22)" - "524288","10.30 (min 10.10, max 10.73)","1.12 (min 0.85, max 1.59)","11.16 (min 10.82, max 11.45)","0.99 (min 0.85, max 1.18)" - "1048576","10.35 (min 10.17, max 10.78)","1.16 (min 0.86, max 1.73)","11.14 (min 10.98, max 11.52)","0.97 (min 0.82, max 1.08)" - "5242880","10.37 (min 10.13, max 10.71)","1.08 (min 0.84, max 1.58)","11.79 (min 11.23, max 12.02)","1.02 (min 0.80, max 1.20)" + "102400","10.66 (min 10.15, max 10.97)","0.70 (min 0.54, max 0.99)","11.02 (min 10.66, max 11.11)","0.76 (min 0.64, max 0.86)" + "262144","10.76 (min 10.52, max 11.00)","0.67 (min 0.45, max 1.02)","11.05 (min 10.81, max 11.20)","0.67 (min 0.62, max 0.74)" + "524288","10.72 (min 10.27, max 11.02)","0.58 (min 0.41, max 0.85)","11.16 (min 10.91, max 11.50)","0.65 (min 0.56, max 0.78)" + "1048576","10.80 (min 10.40, max 11.15)","0.64 (min 0.44, max 0.89)","11.34 (min 10.85, max 12.02)","0.64 (min 0.53, max 0.77)" + "5242880","11.02 (min 10.68, max 11.26)","0.61 (min 0.44, max 0.98)","12.01 (min 11.98, max 12.03)","0.75 (min 0.60, max 0.86)" .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write Raw CPU Load (%)","am62xx_lp_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read Raw CPU Load (%)" - "102400","10.53 (min 10.22, max 10.94)","1.33 (min 1.09, max 1.74)","10.99 (min 10.62, max 11.80)","1.05 (min 0.98, max 1.14)" - "262144","10.33 (min 10.09, max 10.82)","1.37 (min 1.18, max 1.97)","11.06 (min 10.73, max 11.21)","1.22 (min 1.08, max 1.31)" - "524288","10.31 (min 10.10, max 10.74)","1.34 (min 1.12, max 1.80)","11.16 (min 11.05, max 11.50)","1.15 (min 1.08, max 1.26)" - "1048576","10.32 (min 10.08, max 10.72)","1.41 (min 1.15, max 1.81)","11.20 (min 10.84, max 11.50)","1.16 (min 1.09, max 1.25)" - "5242880","10.41 (min 10.15, max 10.69)","1.31 (min 1.11, max 1.92)","11.88 (min 11.50, max 12.02)","1.20 (min 1.10, max 1.26)" + "102400","10.44 (min 10.00, max 10.84)","0.89 (min 0.73, max 1.16)","10.96 (min 10.64, max 11.81)","0.83 (min 0.66, max 0.99)" + "262144","10.41 (min 9.72, max 10.96)","0.82 (min 0.60, max 1.44)","11.04 (min 10.91, max 11.19)","0.81 (min 0.42, max 1.02)" + "524288","10.33 (min 9.37, max 10.79)","0.75 (min 0.60, max 0.97)","11.20 (min 10.99, max 11.50)","0.87 (min 0.75, max 0.97)" + "1048576","10.60 (min 9.85, max 10.97)","0.79 (min 0.59, max 1.26)","11.44 (min 10.96, max 12.00)","0.89 (min 0.80, max 0.94)" + "5242880","10.65 (min 10.25, max 11.15)","0.74 (min 0.59, max 1.04)","12.01 (min 11.99, max 12.02)","0.93 (min 0.86, max 0.98)" .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write Raw CPU Load (%)","am62xxsip_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read Raw CPU Load (%)" - "102400","10.42 (min 8.81, max 10.95)","1.34 (min 0.65, max 2.59)","11.17 (min 10.59, max 11.81)","1.26 (min 0.99, max 1.63)" - "262144","10.46 (min 8.73, max 11.13)","0.97 (min 0.66, max 1.95)","11.14 (min 10.78, max 11.92)","1.14 (min 0.94, max 1.79)" - "524288","10.48 (min 9.36, max 10.84)","0.94 (min 0.71, max 1.40)","11.23 (min 10.91, max 12.00)","1.00 (min 0.87, max 1.18)" - "1048576","10.46 (min 9.35, max 10.92)","0.99 (min 0.74, max 1.31)","11.20 (min 10.83, max 12.01)","1.02 (min 0.85, max 1.19)" - "5242880","10.69 (min 9.03, max 11.30)","0.91 (min 0.61, max 1.29)","11.57 (min 10.87, max 12.03)","1.07 (min 0.85, max 1.26)" + "102400","10.64 (min 9.97, max 10.88)","0.98 (min 0.54, max 1.67)","10.88 (min 10.86, max 10.91)","1.65 (min 1.56, max 1.74)" + "262144","10.64 (min 10.39, max 10.86)","0.90 (min 0.59, max 1.70)","10.92 (min 10.81, max 11.00)","1.12 (min 0.93, max 1.23)" + "524288","10.64 (min 10.41, max 10.90)","0.71 (min 0.46, max 1.37)","11.03 (min 10.94, max 11.08)","1.12 (min 0.93, max 1.27)" + "1048576","11.01 (min 10.55, max 11.38)","1.11 (min 0.49, max 1.90)","11.99 (min 11.96, max 12.01)","1.17 (min 0.92, max 1.34)" + "5242880","11.23 (min 10.86, max 11.33)","1.47 (min 0.51, max 2.99)","12.00 (min 11.99, max 12.02)","1.31 (min 1.00, max 1.57)" The performance numbers were captured using the following: @@ -785,23 +802,24 @@ UBOOT MMCSD FAT .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62xx_sk-fs: Write Throughput (Kbytes/sec)","am62xx_sk-fs: Read Throughput (Kbytes/sec)" - "400000","35466.05 (min 16318.73, max 40960.00)","82755.92 (min 81920.00, max 83591.84)" - "800000","42607.92 (min 36735.43, max 46282.49)","87151.41 (min 86231.58, max 88086.02)" - "1000000","46774.89 (min 42335.92, max 49201.20)","89653.03 (min 89530.05, max 90021.98)" + "400000","36246.41 (min 21787.23, max 40554.46)","82198.64 (min 81920.00, max 83591.84)" + "800000","44588.63 (min 35310.34, max 56109.59)","86846.44 (min 86231.58, max 88086.02)" + "1000000","48914.57 (min 44521.74, max 51684.54)","89369.63 (min 88562.16, max 90021.98)" + .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62xx_lp_sk-fs: Write Throughput (Kbytes/sec)","am62xx_lp_sk-fs: Read Throughput (Kbytes/sec)" - "400000","36399.32 (min 30567.16, max 39766.99)","82636.50 (min 81920.00, max 83591.84)" - "800000","41351.40 (min 35008.55, max 46545.45)","87148.94" - "1000000","48424.06 (min 44281.08, max 50103.98)","89670.60 (min 89530.05, max 90021.98)" + "400000","36781.47 (min 29049.65, max 40960.00)","81952.15 (min 78769.23, max 83591.84)" + "800000","43935.80 (min 30228.78, max 50257.67)","86035.96 (min 79533.98, max 88086.02)" + "1000000","47180.72 (min 42010.26, max 53368.08)","88994.57 (min 85333.33, max 90021.98)" .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62xxsip_sk-fs: Write Throughput (Kbytes/sec)","am62xxsip_sk-fs: Read Throughput (Kbytes/sec)" - "400000","31694.74 (min 18875.58, max 39384.62)","81820.63 (min 78769.23, max 83591.84)" - "800000","37983.33 (min 20897.96, max 45765.36)","86655.57 (min 85333.33, max 88086.02)" - "1000000","41131.43 (min 21005.13, max 49053.89)","89444.00 (min 88562.16, max 90021.98)" + "400000","27632.53 (min 16253.97, max 40554.46)","81427.02 (min 78769.23, max 83591.84)" + "800000","37337.86 (min 20480.00, max 46022.47)","86543.74 (min 85333.33, max 87148.94)" + "1000000","36834.85 (min 20152.52, max 47906.43)","89371.40 (min 88562.16, max 90021.98)" The performance numbers were captured using the following: @@ -809,93 +827,94 @@ The performance numbers were captured using the following: | -USB Driver ----------- +USB +=== USB Device Controller -^^^^^^^^^^^^^^^^^^^^^ +--------------------- .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)","am62xxsip_sk-fs: Throughput (MB/sec)" - "150","33.11 (min 30.30, max 34.50)","43.23 (min 42.60, max 44.20)","44.00 (min 43.60, max 44.40)" + "150","43.53 (min 43.30, max 43.70)","43.27 (min 42.70, max 43.60)","43.90 (min 43.80, max 44.00)" .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)","am62xxsip_sk-fs: Throughput (MB/sec)" - "150","29.70 (min 26.70, max 32.00)","41.21 (min 40.30, max 42.90)","40.01 (min 37.50, max 42.30)" + "150","39.40 (min 39.10, max 39.80)","41.35 (min 40.90, max 41.90)","39.70 (min 38.70, max 40.70)" | -CRYPTO Driver -------------- +CRYPTO +====== OpenSSL Performance -^^^^^^^^^^^^^^^^^^^ +------------------- .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am62xx_lp_sk-fs: throughput (KBytes/Sec)","am62xx_sk-fs: throughput (KBytes/Sec)","am62xxsip_sk-fs: throughput (KBytes/Sec)" - "aes-128-cbc","1024","20907.99 (min 19976.19, max 22306.47)","23639.08 (min 21773.65, max 24485.89)","22077.10 (min 21114.88, max 23049.56)" - "aes-128-cbc","16","364.55 (min 352.28, max 374.50)","387.41 (min 379.85, max 419.58)","380.12 (min 370.18, max 392.97)" - "aes-128-cbc","16384","116402.86 (min 113541.12, max 119870.81)","120229.89 (min 115938.65, max 131110.23)","122670.42 (min 117544.28, max 127505.75)" - "aes-128-cbc","256","5982.09 (min 5769.30, max 6152.11)","6286.31 (min 6142.29, max 7050.92)","6368.79 (min 6144.85, max 6681.17)" - "aes-128-cbc","64","1497.61 (min 1441.62, max 1537.66)","1584.44 (min 1541.50, max 1818.71)","1617.02 (min 1536.41, max 1708.99)" - "aes-128-cbc","8192","88300.89 (min 85458.94, max 91952.47)","93108.22 (min 88402.60, max 100466.69)","92829.70 (min 89186.30, max 96419.84)" - "aes-128-ecb","1024","21740.50 (min 20489.90, max 22819.16)","24123.86 (min 22792.87, max 25001.98)","22632.58 (min 21597.87, max 23622.31)" - "aes-128-ecb","16","370.24 (min 357.28, max 381.93)","390.43 (min 383.62, max 431.56)","386.75 (min 377.13, max 398.98)" - "aes-128-ecb","16384","120206.68 (min 117850.11, max 123016.53)","124764.84 (min 120045.57, max 136303.96)","126950.06 (min 121247.06, max 131765.59)" - "aes-128-ecb","256","5999.26 (min 5773.74, max 6163.88)","6343.30 (min 6170.97, max 7289.86)","6470.59 (min 6142.12, max 6830.59)" - "aes-128-ecb","64","1505.50 (min 1446.49, max 1541.23)","1624.92 (min 1547.71, max 1877.35)","1642.30 (min 1537.45, max 1747.48)" - "aes-128-ecb","8192","90868.39 (min 88468.14, max 92310.19)","97160.87 (min 91930.62, max 103997.44)","95978.84 (min 92334.76, max 99461.80)" - "aes-192-cbc","1024","20752.00 (min 19784.02, max 22273.71)","23517.06 (min 21488.64, max 24410.45)","21847.89 (min 20898.82, max 22822.57)" - "aes-192-cbc","16","365.52 (min 353.82, max 376.50)","388.34 (min 381.74, max 421.03)","379.18 (min 369.00, max 387.85)" - "aes-192-cbc","16384","110072.49 (min 107440.81, max 112295.94)","113373.87 (min 109625.34, max 124081.49)","116275.20 (min 112104.79, max 120176.64)" - "aes-192-cbc","256","5979.54 (min 5767.85, max 6141.44)","6288.97 (min 6152.28, max 7104.60)","6333.77 (min 6110.21, max 6583.64)" - "aes-192-cbc","64","1498.07 (min 1442.62, max 1538.50)","1588.33 (min 1542.44, max 1838.59)","1615.22 (min 1536.38, max 1698.15)" - "aes-192-cbc","8192","84092.59 (min 81810.77, max 86608.55)","88387.58 (min 84429.48, max 96318.81)","89152.51 (min 85595.48, max 92646.06)" - "aes-192-ecb","1024","21486.68 (min 20369.07, max 22924.63)","24089.77 (min 22578.52, max 25076.74)","22564.44 (min 21724.16, max 23545.86)" - "aes-192-ecb","16","370.41 (min 358.32, max 382.93)","390.16 (min 383.19, max 431.48)","387.65 (min 379.52, max 398.96)" - "aes-192-ecb","16384","114104.32 (min 112219.48, max 117456.90)","118111.57 (min 113617.58, max 129111.38)","120551.42 (min 114524.16, max 125572.44)" - "aes-192-ecb","256","5993.78 (min 5765.89, max 6156.20)","6344.05 (min 6169.69, max 7315.54)","6475.15 (min 6148.18, max 6893.74)" - "aes-192-ecb","64","1502.55 (min 1446.85, max 1541.78)","1627.19 (min 1546.43, max 1887.98)","1648.31 (min 1540.27, max 1742.19)" - "aes-192-ecb","8192","87736.66 (min 85093.03, max 91247.96)","92213.59 (min 87569.75, max 99887.79)","92321.11 (min 88678.40, max 95783.59)" - "aes-256-cbc","1024","20581.29 (min 19701.76, max 21914.28)","23211.82 (min 21145.26, max 24292.69)","21607.81 (min 20540.07, max 22718.46)" - "aes-256-cbc","16","365.27 (min 351.82, max 377.43)","388.41 (min 383.25, max 419.81)","379.23 (min 368.19, max 389.95)" - "aes-256-cbc","16384","102187.01 (min 99074.05, max 105633.11)","106252.97 (min 102176.09, max 115627.35)","107868.16 (min 102318.08, max 112678.23)" - "aes-256-cbc","256","5974.42 (min 5769.22, max 6132.31)","6279.75 (min 6146.82, max 7036.33)","6326.14 (min 6087.85, max 6636.97)" - "aes-256-cbc","64","1497.24 (min 1443.63, max 1538.05)","1587.18 (min 1540.95, max 1830.78)","1607.81 (min 1514.67, max 1704.75)" - "aes-256-cbc","8192","80066.90 (min 78809.77, max 81939.11)","84246.19 (min 80595.63, max 91176.96)","84435.29 (min 80655.70, max 87881.05)" - "aes-256-ecb","1024","21426.18 (min 20330.15, max 22955.35)","23968.47 (min 22278.83, max 24843.26)","22376.92 (min 21511.85, max 23244.80)" - "aes-256-ecb","16","370.62 (min 359.25, max 381.87)","390.07 (min 382.78, max 431.39)","387.06 (min 379.45, max 397.67)" - "aes-256-ecb","16384","107402.58 (min 105185.28, max 109685.42)","110622.72 (min 107080.36, max 120558.93)","112945.83 (min 108505.77, max 117036.37)" - "aes-256-ecb","256","5992.09 (min 5767.00, max 6159.53)","6339.20 (min 6167.81, max 7293.70)","6456.95 (min 6145.62, max 6806.61)" - "aes-256-ecb","64","1502.26 (min 1448.23, max 1542.25)","1624.00 (min 1546.86, max 1885.10)","1648.13 (min 1540.82, max 1750.91)" - "aes-256-ecb","8192","83537.24 (min 81562.28, max 84948.31)","87660.54 (min 84374.87, max 95049.05)","87739.05 (min 84350.29, max 90947.58)" - "sha256","1024","31667.97 (min 30793.05, max 32121.86)","36792.92 (min 36691.63, max 37363.03)","32547.50 (min 31477.76, max 33580.71)" - "sha256","16","535.08 (min 517.10, max 544.56)","581.78 (min 574.87, max 626.29)","546.96 (min 528.06, max 563.91)" - "sha256","16384","259375.10 (min 256791.89, max 261860.01)","294220.46 (min 292962.30, max 298920.62)","263269.72 (min 260674.90, max 268413.61)" - "sha256","256","8403.27 (min 8144.30, max 8599.30)","9269.37 (min 9191.00, max 9772.97)","8519.69 (min 8210.35, max 8809.05)" - "sha256","64","2118.85 (min 2050.79, max 2152.09)","2321.29 (min 2297.30, max 2466.28)","2161.82 (min 2093.80, max 2229.46)" - "sha256","8192","175455.91 (min 172878.51, max 179773.44)","197163.69 (min 196318.55, max 202429.78)","176976.21 (min 172785.66, max 181379.07)" - "sha512","1024","22679.64 (min 22215.00, max 23011.33)","24744.92 (min 24600.58, max 25694.21)","22754.13 (min 22400.00, max 23131.14)" - "sha512","16","522.33 (min 501.00, max 537.31)","577.81 (min 573.24, max 601.51)","529.39 (min 512.75, max 543.56)" - "sha512","16384","60596.22 (min 60304.04, max 60926.63)","67861.16 (min 67671.38, max 68299.43)","60706.82 (min 60342.27, max 61063.17)" - "sha512","256","7290.51 (min 7085.40, max 7548.84)","8234.50 (min 8163.75, max 8537.43)","7490.46 (min 7222.87, max 7709.53)" - "sha512","64","2095.92 (min 2021.93, max 2149.91)","2311.92 (min 2293.89, max 2405.40)","2119.25 (min 2058.05, max 2178.41)" - "sha512","8192","54249.47 (min 53824.17, max 54613.33)","60518.40 (min 60325.89, max 61390.85)","54479.87 (min 53963.43, max 54910.98)" + "aes-128-cbc","1024","21742.25","24533.33","24039.08" + "aes-128-cbc","16","378.80","385.16","383.05" + "aes-128-cbc","16384","117959.34","120673.62","122934.61" + "aes-128-cbc","256","6155.09","6233.26","6222.34" + "aes-128-cbc","64","1538.09","1639.94","1591.89" + "aes-128-cbc","8192","89830.74","94538.41","97886.21" + "aes-128-ecb","1024","22719.15","24592.38","24507.05" + "aes-128-ecb","16","383.20","386.55","386.23" + "aes-128-ecb","16384","120640.85","125042.69","129351.68" + "aes-128-ecb","256","6168.58","6537.22","6389.93" + "aes-128-ecb","64","1545.64","1729.02","1667.37" + "aes-128-ecb","8192","92547.75","98254.85","98358.61" + "aes-192-cbc","1024","21413.89","24498.52","24373.93" + "aes-192-cbc","16","381.31","384.91","385.23" + "aes-192-cbc","16384","110351.70","115228.67","115769.34" + "aes-192-cbc","256","6150.74","6228.48","6216.87" + "aes-192-cbc","64","1540.44","1636.27","1615.96" + "aes-192-cbc","8192","84582.40","90395.99","90745.51" + "aes-192-ecb","1024","22448.47","24605.70","24558.59" + "aes-192-ecb","16","380.79","386.39","386.22" + "aes-192-ecb","16384","114098.18","118035.80","121055.91" + "aes-192-ecb","256","6169.51","6392.92","6323.97" + "aes-192-ecb","64","1546.37","1708.93","1698.71" + "aes-192-ecb","8192","88380.76","92755.29","96624.64" + "aes-256-cbc","1024","21324.46","24482.13","24067.75" + "aes-256-cbc","16","381.35","385.14","384.90" + "aes-256-cbc","16384","101111.13","106577.92","107533.65" + "aes-256-cbc","256","6143.40","6218.84","6203.22" + "aes-256-cbc","64","1539.43","1628.82","1567.98" + "aes-256-cbc","8192","81002.50","84445.87","85393.41" + "aes-256-ecb","1024","22270.63","24578.39","24514.56" + "aes-256-ecb","16","382.64","386.23","386.61" + "aes-256-ecb","16384","107457.19","110526.46","111771.65" + "aes-256-ecb","256","6163.37","6369.54","6258.69" + "aes-256-ecb","64","1543.34","1707.82","1700.89" + "aes-256-ecb","8192","84467.71","88790.36","89587.71" + "sha256","1024","24685.57","28421.80","28115.29" + "sha256","16","387.77","478.60","478.12" + "sha256","16384","237305.86","262215.00","261783.55" + "sha256","256","6193.58","7474.60","7386.62" + "sha256","64","1550.25","1890.03","1877.25" + "sha256","8192","156396.20","173823.32","172504.41" + "sha512","1024","22674.09","24628.57","24615.59" + "sha512","16","386.84","468.11","462.11" + "sha512","16384","97610.41","108729.69","108882.60" + "sha512","256","6166.53","6850.47","6755.50" + "sha512","64","1547.43","1872.62","1851.22" + "sha512","8192","78539.43","89071.62","88926.89" + .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am62xx_lp_sk-fs: CPU Load","am62xx_sk-fs: CPU Load","am62xxsip_sk-fs: CPU Load" - "aes-128-cbc","36.63 (min 34.00, max 38.00)","36.88 (min 35.00, max 38.00)","36.38 (min 35.00, max 37.00)" - "aes-128-ecb","38.75 (min 36.00, max 40.00)","39.13 (min 37.00, max 40.00)","38.13 (min 36.00, max 39.00)" - "aes-192-cbc","37.00 (min 35.00, max 39.00)","37.50 (min 36.00, max 39.00)","35.75 (min 32.00, max 38.00)" - "aes-192-ecb","38.50 (min 36.00, max 40.00)","38.38 (min 37.00, max 40.00)","37.88 (min 36.00, max 39.00)" - "aes-256-cbc","36.88 (min 35.00, max 38.00)","36.75 (min 35.00, max 38.00)","36.13 (min 34.00, max 38.00)" - "aes-256-ecb","38.25 (min 36.00, max 40.00)","38.13 (min 37.00, max 39.00)","37.38 (min 35.00, max 39.00)" - "sha256","93.75 (min 93.00, max 94.00)","94.38 (min 93.00, max 95.00)","94.50 (min 93.00, max 95.00)" - "sha512","93.38 (min 92.00, max 94.00)","94.75 (min 94.00, max 95.00)","94.63 (min 94.00, max 95.00)" + "aes-128-cbc","38.00","38.00","38.00" + "aes-128-ecb","40.00","39.00","39.00" + "aes-192-cbc","39.00","38.00","38.00" + "aes-192-ecb","40.00","38.00","39.00" + "aes-256-cbc","38.00","38.00","38.00" + "aes-256-ecb","39.00","38.00","37.00" + "sha256","95.00","95.00","94.00" + "sha512","95.00","95.00","94.00" Listed for each algorithm are the code snippets used to run each benchmark test. @@ -905,73 +924,13 @@ Listed for each algorithm are the code snippets used to run each time -v openssl speed -elapsed -evp aes-128-cbc IPSec Software Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^ +-------------------------- .. csv-table:: IPSec Software Performance :header: "Algorithm","am62xx_sk-fs: Throughput (Mbps)","am62xx_sk-fs: Packets/Sec","am62xx_sk-fs: CPU Load","am62xxsip_sk-fs: Throughput (Mbps)","am62xxsip_sk-fs: Packets/Sec","am62xxsip_sk-fs: CPU Load" - "aes128","121.71 (min 2.30, max 305.50)","10.57 (min 0.00, max 27.00)","48.76 (min 28.52, max 54.93)","135.97 (min 4.10, max 338.50)","11.71 (min 0.00, max 30.00)","52.73 (min 47.19, max 58.90)" - "aes192","75.94 (min 0.20, max 228.80)","6.43 (min 0.00, max 20.00)","42.84 (min 28.19, max 50.88)","2.40 (min 2.20, max 2.60)","0.00","85.06 (min 83.00, max 87.11)" - "aes256","188.42 (min 1.30, max 302.20)","16.33 (min 0.00, max 26.00)","48.78 (min 28.23, max 54.44)","164.90 (min 4.80, max 331.20)","14.33 (min 0.00, max 29.00)","53.12 (min 28.47, max 65.09)" - -Low Power Performance -------------------------- - -Power Performance -^^^^^^^^^^^^^^^^^ - -.. csv-table:: Deep Sleep Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core","0.85","9.59" - "vddr_core","0.85","n/a" - "soc_dvdd_3v3","3.30","7.71" - "soc_dvdd_1v8","1.80","0.54" - "vdda_1v8","1.80","1.89" - "vdd_ddr4","1.10","8.49" - "Total"," ","28.21" + "aes128","77.65 (min 73.60, max 81.70)","6.50 (min 6.00, max 7.00)","51.08 (min 50.78, max 51.38)","80.50","7.00","51.43" + "aes192","0.60","0.00","56.22" + "aes256","130.00 (min 53.40, max 206.60)","11.00 (min 4.00, max 18.00)","39.11 (min 27.23, max 50.98)","88.60 (min 0.40, max 205.40)","7.67 (min 0.00, max 18.00)","43.07 (min 27.21, max 51.38)" -.. csv-table:: MCU Only Power Performance - :header: "Rail name","Rail voltage(V)","Power (mW)" - - "vdd_core","0.85","110.78" - "vddr_core","0.85","n/a" - "soc_dvdd_3v3","3.30","13.81" - "soc_dvdd_1v8","1.80","0.50" - "vdda_1v8","1.80","11.35" - "vdd_ddr4","1.10","8.79" - "Total"," ","145.23" - -Partial I/O Data -- All voltage rails were measured to be near 0V - -.. note:: - - The measurements shown are from an SK-AM62B-P1. Results may vary based off of the board variant being used. - -Further optimizations are possible for these low power modes. Please refer to the AM62x Power Consumption App Note (https://www.ti.com/lit/pdf/spradg1) - -Resume Latency Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^ - -.. csv-table:: LPM Resume Latency Performance - :header: "Low Power Mode","Total Resume Latency (ms)" - - "Deep Sleep", "131.50" - "MCU Only", "84.89" - -The performance numbers are measured without the Linux printk logs. To remove the -Linux printk logs, run the following commands in the terminal: - -.. code:: console - - # Detach kernel serial console - consoles=$(find /sys/bus/platform/devices/*.serial/ -name console) - for console in ${consoles}; do - echo -n N > ${console} - done - - -.. note:: - - The measurements shown are from using the default SDK with no extra optimizations. +| diff --git a/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst index 65ea7a22b..415819b2a 100644 --- a/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst @@ -1,9 +1,10 @@ -======================================= - RT-linux 11.02.08.02 Performance Guide -======================================= +################################### +RT-linux 12.00.00 Performance Guide +################################### -.. rubric:: **Read This First** - :name: read-this-first-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -20,8 +21,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-rt-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -29,62 +31,13 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-rt-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ | +***************** System Benchmarks ------------------ - -Stress-ng and Cyclic Test - -.. _RT-linux-performance: - -stress-ng (next-generation) will stress test a embedded platform in various selectable ways. -It was designed to exercise various physical subsystems as well as the various -operating system kernel interfaces. stress-ng can also measure test throughput rates; -this can be useful to observe performance changes across different operating system or types of hardware. - -Cyclictest is most commonly used for benchmarking RT systems. -It is one of the most frequently used tools for evaluating the relative performance of real-time systems. -Some performance tests which use Cyclictest are System benchmarking, Latency debugging with tracing and -approximating application performance. - -Test commands used for running stress-ng and cyclictest together - -.. code:: console - - stress-ng --cpu-method=all -c 4 & - cyclictest -m -Sp80 -D6h -h400 -i200 -M -q +***************** -The following summarizes the latencies observed using the yocto based -default SDK image using the SK-AM62B-P1_ reference board: - -.. _SK-AM62B-P1: https://www.ti.com/tool/SK-AM62B-P1 - -.. note:: - - Using the OP-TEE TRNG driver can impact this benchmark's performance due to - frequent context switching between Normal World (Linux) and Secure World (OP-TEE), - that occurs when the kernel's hardware random number generator interface - (hwrng) requests entropy from the secure TRNG to replenish the Linux entropy - pool. - - The Linux TRNG driver can mitigate these latency spikes. This involves - enabling the Pseudo RNG driver in OP-TEE as documented in the Foundational - Components section: :ref:`building-optee-with-prng`, and enabling the RNG - node in the Linux kernel device tree. This way the HW TRNG is accessed from - the kernel itself. - -.. csv-table:: - :header: "Latencies","CPU 0","CPU 1","CPU 2","CPU 3" - - "Minimum (usec)","5","5","5","5" - "Average (usec)","6","6","6","6" - "Maximum (usec)","66","41","39","31" - -.. image:: img/rt-cpu-method-all-latency-histogram.png +| diff --git a/source/devices/AM64X/linux/Linux_Performance_Guide.rst b/source/devices/AM64X/linux/Linux_Performance_Guide.rst index 2b2772afa..7d53d8a38 100644 --- a/source/devices/AM64X/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM64X/linux/Linux_Performance_Guide.rst @@ -1,8 +1,10 @@ -==================================== - Linux 11.02.08.02 Performance Guide -==================================== +################################# + Linux 12.00.00 Performance Guide +################################# -.. rubric:: **Read This First** +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -15,8 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -24,8 +27,13 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ + +| + +***************** +System Benchmarks +***************** + +| diff --git a/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst index 930291f94..24b0c1f7c 100644 --- a/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst @@ -1,9 +1,10 @@ -====================================== -RT-linux 11.02.08.02 Performance Guide -====================================== +################################### +RT-linux 12.00.00 Performance Guide +################################### -.. rubric:: **Read This First** - :name: read-this-first-rt-kernel-perf-guide +*************** +Read This First +*************** **All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.** @@ -16,8 +17,9 @@ following Evaluation Modules unless otherwise specified.** Table: Evaluation Modules -.. rubric:: About This Manual - :name: about-this-manual-rt-kernel-perf-guide +***************** +About This Manual +***************** This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be @@ -25,66 +27,20 @@ used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release. -.. rubric:: If You Need Assistance - :name: if-you-need-assistance-rt-kernel-perf-guide - For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ | +***************** System Benchmarks ------------------ - -Stress-ng and Cyclic Test -^^^^^^^^^^^^^^^^^^^^^^^^^ - -.. _RT-linux-performance: - -stress-ng (next-generation) will stress test a embedded platform in -various selectable ways. It was designed to exercise various physical -subsystems as well as the various operating system kernel interfaces. -stress-ng can also measure test throughput rates; this can be useful to -observe performance changes across different operating system or types -of hardware. - -Cyclictest is most commonly used for benchmarking RT systems. It is one -of the most frequently used tools for evaluating the relative -performance of real-time systems. Some performance tests which use -Cyclictest are System benchmarking, Latency debugging with tracing and --approximating application performance. - -Test commands used for running stress-ng and cyclictest together - -.. code:: console - - stress-ng --cpu-method=all -c 4 & - cyclictest -m -Sp80 -D6h -h400 -i200 -M -q +***************** -The following summarizes the latencies observed using the yocto based -default SDK image: - -.. note:: - - A known issue in this SDK release is affecting this benchmark. - Using OP-TEE's PRNG drivers rather than the hardware accelerated TRNG - drivers restores the context switch latencies to the values you see - here. - - More information on switching to the PRNG drivers can be found in the - Foundational Components section, here :ref:`building-optee-with-prng` - -.. csv-table:: - :header: "Latencies","CPU 0","CPU 1" - - "Minimum (usec)","6","6" - "Average (usec)","9","9" - "Maximum (usec)","58","73" - -.. image:: img/rt-cpu-method-all-latency-histogram.png +| LMBench -^^^^^^^ +======= + LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at @@ -111,124 +67,125 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "af_unix_sock_stream_latency (microsec)","42.81 (min 37.94, max 47.17)" - "af_unix_socket_stream_bandwidth (mb\s)","541.45 (min 530.54, max 556.53)" - "bw_file_rd-io-1mb (mb/s)","875.51 (min 850.20, max 890.31)" - "bw_file_rd-o2c-1mb (mb/s)","491.77 (min 456.00, max 508.82)" - "bw_mem-bcopy-16mb (mb/s)","969.07 (min 937.70, max 1007.87)" - "bw_mem-bcopy-1mb (mb/s)","947.26 (min 915.75, max 980.01)" - "bw_mem-bcopy-2mb (mb/s)","934.26 (min 902.53, max 970.87)" - "bw_mem-bcopy-4mb (mb/s)","949.68 (min 911.68, max 983.28)" - "bw_mem-bcopy-8mb (mb/s)","976.43 (min 911.99, max 1010.61)" - "bw_mem-bzero-16mb (mb/s)","2117.38 (min 2116.68, max 2118.08)" - "bw_mem-bzero-1mb (mb/s)","1532.87 (min 915.75, max 2119.87)" - "bw_mem-bzero-2mb (mb/s)","1524.64 (min 902.53, max 2117.90)" - "bw_mem-bzero-4mb (mb/s)","1532.55 (min 911.68, max 2118.64)" - "bw_mem-bzero-8mb (mb/s)","1546.42 (min 911.99, max 2118.36)" - "bw_mem-cp-16mb (mb/s)","592.63 (min 576.54, max 600.85)" - "bw_mem-cp-1mb (mb/s)","1585.65 (min 543.97, max 2718.45)" - "bw_mem-cp-2mb (mb/s)","1432.40 (min 524.11, max 2363.14)" - "bw_mem-cp-4mb (mb/s)","1376.53 (min 523.49, max 2214.43)" - "bw_mem-cp-8mb (mb/s)","1352.70 (min 523.01, max 2157.79)" - "bw_mem-fcp-16mb (mb/s)","1051.58 (min 1000.13, max 1099.43)" - "bw_mem-fcp-1mb (mb/s)","1591.05 (min 1028.10, max 2119.87)" - "bw_mem-fcp-2mb (mb/s)","1567.15 (min 987.33, max 2117.90)" - "bw_mem-fcp-4mb (mb/s)","1565.96 (min 982.08, max 2118.64)" - "bw_mem-fcp-8mb (mb/s)","1571.64 (min 971.46, max 2118.36)" - "bw_mem-frd-16mb (mb/s)","1317.14 (min 1282.36, max 1364.84)" - "bw_mem-frd-1mb (mb/s)","1170.35 (min 1028.10, max 1305.24)" - "bw_mem-frd-2mb (mb/s)","1167.91 (min 987.33, max 1373.86)" - "bw_mem-frd-4mb (mb/s)","1161.96 (min 982.08, max 1330.23)" - "bw_mem-frd-8mb (mb/s)","1175.62 (min 971.46, max 1385.76)" - "bw_mem-fwr-16mb (mb/s)","2122.36 (min 2106.37, max 2131.91)" - "bw_mem-fwr-1mb (mb/s)","1933.56 (min 1216.12, max 2718.45)" - "bw_mem-fwr-2mb (mb/s)","1822.22 (min 1261.83, max 2363.14)" - "bw_mem-fwr-4mb (mb/s)","1754.10 (min 1265.82, max 2214.43)" - "bw_mem-fwr-8mb (mb/s)","1736.68 (min 1236.09, max 2157.79)" - "bw_mem-rd-16mb (mb/s)","1366.15 (min 1338.24, max 1382.65)" - "bw_mem-rd-1mb (mb/s)","1100.78 (min 793.27, max 1403.09)" - "bw_mem-rd-2mb (mb/s)","1101.55 (min 801.18, max 1400.81)" - "bw_mem-rd-4mb (mb/s)","1105.82 (min 824.91, max 1399.09)" - "bw_mem-rd-8mb (mb/s)","1125.63 (min 861.05, max 1393.73)" - "bw_mem-rdwr-16mb (mb/s)","856.58 (min 844.73, max 870.65)" - "bw_mem-rdwr-1mb (mb/s)","693.30 (min 543.97, max 838.64)" - "bw_mem-rdwr-2mb (mb/s)","671.55 (min 524.11, max 834.84)" - "bw_mem-rdwr-4mb (mb/s)","684.01 (min 523.49, max 844.77)" - "bw_mem-rdwr-8mb (mb/s)","701.39 (min 523.01, max 857.08)" - "bw_mem-wr-16mb (mb/s)","893.90 (min 886.48, max 899.79)" - "bw_mem-wr-1mb (mb/s)","819.90 (min 754.86, max 847.17)" - "bw_mem-wr-2mb (mb/s)","814.11 (min 777.91, max 836.35)" - "bw_mem-wr-4mb (mb/s)","828.96 (min 793.41, max 876.33)" - "bw_mem-wr-8mb (mb/s)","864.35 (min 836.65, max 897.06)" - "bw_mmap_rd-mo-1mb (mb/s)","1320.67 (min 1272.50, max 1351.35)" - "bw_mmap_rd-o2c-1mb (mb/s)","474.18 (min 454.27, max 483.64)" - "bw_pipe (mb/s)","540.02 (min 518.86, max 558.01)" - "bw_unix (mb/s)","541.45 (min 530.54, max 556.53)" - "lat_connect (us)","75.76 (min 74.84, max 78.74)" - "lat_ctx-2-128k (us)","10.34 (min 8.76, max 11.41)" - "lat_ctx-2-256k (us)","22.65 (min 15.21, max 28.60)" - "lat_ctx-4-128k (us)","11.46 (min 10.21, max 12.85)" - "lat_ctx-4-256k (us)","10.78 (min 0.00, max 18.44)" - "lat_fs-0k (num_files)","198.38 (min 194.00, max 203.00)" - "lat_fs-10k (num_files)","88.38 (min 82.00, max 103.00)" - "lat_fs-1k (num_files)","132.88 (min 129.00, max 138.00)" - "lat_fs-4k (num_files)","118.50 (min 107.00, max 126.00)" - "lat_mem_rd-stride128-sz1000k (ns)","47.38 (min 46.80, max 48.23)" - "lat_mem_rd-stride128-sz125k (ns)","7.85 (min 7.80, max 7.97)" - "lat_mem_rd-stride128-sz250k (ns)","12.09 (min 10.07, max 14.63)" - "lat_mem_rd-stride128-sz31k (ns)","5.25 (min 3.07, max 5.90)" + "af_unix_sock_stream_latency (microsec)","29.27 (min 24.87, max 39.15)" + "af_unix_socket_stream_bandwidth (mb\s)","558.67 (min 544.62, max 577.21)" + "bw_file_rd-io-1mb (mb/s)","842.86 (min 820.61, max 860.44)" + "bw_file_rd-o2c-1mb (mb/s)","494.65 (min 478.77, max 505.56)" + "bw_mem-bcopy-16mb (mb/s)","949.99 (min 921.34, max 973.47)" + "bw_mem-bcopy-1mb (mb/s)","943.11 (min 906.76, max 975.93)" + "bw_mem-bcopy-2mb (mb/s)","933.38 (min 918.70, max 958.47)" + "bw_mem-bcopy-4mb (mb/s)","960.36 (min 924.21, max 978.47)" + "bw_mem-bcopy-8mb (mb/s)","954.73 (min 901.51, max 986.19)" + "bw_mem-bzero-16mb (mb/s)","2117.57 (min 2115.56, max 2118.36)" + "bw_mem-bzero-1mb (mb/s)","1530.94 (min 906.76, max 2120.52)" + "bw_mem-bzero-2mb (mb/s)","1525.61 (min 918.70, max 2118.64)" + "bw_mem-bzero-4mb (mb/s)","1538.69 (min 924.21, max 2117.90)" + "bw_mem-bzero-8mb (mb/s)","1536.19 (min 901.51, max 2118.92)" + "bw_mem-cp-16mb (mb/s)","561.34 (min 542.93, max 581.71)" + "bw_mem-cp-1mb (mb/s)","1606.01 (min 557.72, max 2718.38)" + "bw_mem-cp-2mb (mb/s)","1439.68 (min 534.62, max 2352.94)" + "bw_mem-cp-4mb (mb/s)","1383.16 (min 524.25, max 2227.17)" + "bw_mem-cp-8mb (mb/s)","1359.16 (min 537.42, max 2167.43)" + "bw_mem-fcp-16mb (mb/s)","1043.67 (min 1008.64, max 1082.84)" + "bw_mem-fcp-1mb (mb/s)","1605.66 (min 1081.08, max 2120.52)" + "bw_mem-fcp-2mb (mb/s)","1568.44 (min 966.81, max 2118.64)" + "bw_mem-fcp-4mb (mb/s)","1582.47 (min 951.70, max 2117.90)" + "bw_mem-fcp-8mb (mb/s)","1571.60 (min 959.23, max 2118.92)" + "bw_mem-frd-16mb (mb/s)","1311.63 (min 1279.49, max 1329.57)" + "bw_mem-frd-1mb (mb/s)","1200.07 (min 1081.08, max 1331.78)" + "bw_mem-frd-2mb (mb/s)","1168.93 (min 966.81, max 1370.33)" + "bw_mem-frd-4mb (mb/s)","1181.54 (min 951.70, max 1369.63)" + "bw_mem-frd-8mb (mb/s)","1170.23 (min 959.23, max 1367.05)" + "bw_mem-fwr-16mb (mb/s)","2128.04 (min 2125.96, max 2129.93)" + "bw_mem-fwr-1mb (mb/s)","1965.95 (min 1273.19, max 2718.38)" + "bw_mem-fwr-2mb (mb/s)","1819.17 (min 1265.22, max 2352.94)" + "bw_mem-fwr-4mb (mb/s)","1760.80 (min 1275.31, max 2227.17)" + "bw_mem-fwr-8mb (mb/s)","1737.99 (min 1241.85, max 2167.43)" + "bw_mem-rd-16mb (mb/s)","1332.85 (min 1270.14, max 1382.89)" + "bw_mem-rd-1mb (mb/s)","1102.66 (min 850.77, max 1419.88)" + "bw_mem-rd-2mb (mb/s)","1086.05 (min 821.81, max 1365.89)" + "bw_mem-rd-4mb (mb/s)","1104.47 (min 856.62, max 1380.74)" + "bw_mem-rd-8mb (mb/s)","1116.97 (min 863.84, max 1397.38)" + "bw_mem-rdwr-16mb (mb/s)","878.96 (min 870.37, max 884.08)" + "bw_mem-rdwr-1mb (mb/s)","698.17 (min 557.72, max 847.46)" + "bw_mem-rdwr-2mb (mb/s)","684.70 (min 534.62, max 828.16)" + "bw_mem-rdwr-4mb (mb/s)","697.23 (min 524.25, max 855.25)" + "bw_mem-rdwr-8mb (mb/s)","713.67 (min 537.42, max 878.54)" + "bw_mem-wr-16mb (mb/s)","898.11 (min 890.57, max 901.46)" + "bw_mem-wr-1mb (mb/s)","835.61 (min 768.34, max 890.31)" + "bw_mem-wr-2mb (mb/s)","823.57 (min 790.62, max 852.64)" + "bw_mem-wr-4mb (mb/s)","850.08 (min 805.07, max 878.64)" + "bw_mem-wr-8mb (mb/s)","875.81 (min 858.74, max 897.87)" + "bw_mmap_rd-mo-1mb (mb/s)","1288.67 (min 1251.79, max 1327.27)" + "bw_mmap_rd-o2c-1mb (mb/s)","503.15 (min 477.40, max 532.86)" + "bw_pipe (mb/s)","559.24 (min 519.56, max 584.19)" + "bw_unix (mb/s)","558.67 (min 544.62, max 577.21)" + "lat_connect (us)","69.65 (min 68.59, max 71.76)" + "lat_ctx-2-128k (us)","17.41 (min 8.52, max 35.96)" + "lat_ctx-2-256k (us)","59.43 (min 18.26, max 137.82)" + "lat_ctx-4-128k (us)","33.39 (min 10.30, max 79.35)" + "lat_ctx-4-256k (us)","57.09 (min 5.26, max 145.73)" + "lat_fs-0k (num_files)","222.33 (min 204.00, max 253.00)" + "lat_fs-10k (num_files)","97.33 (min 88.00, max 117.00)" + "lat_fs-1k (num_files)","138.83 (min 128.00, max 152.00)" + "lat_fs-4k (num_files)","129.17 (min 126.00, max 135.00)" + "lat_mem_rd-stride128-sz1000k (ns)","46.99 (min 46.70, max 48.08)" + "lat_mem_rd-stride128-sz125k (ns)","7.81 (min 7.80, max 7.85)" + "lat_mem_rd-stride128-sz250k (ns)","13.11 (min 10.29, max 17.26)" + "lat_mem_rd-stride128-sz31k (ns)","4.69 (min 3.06, max 5.89)" "lat_mem_rd-stride128-sz50 (ns)","3.02" - "lat_mem_rd-stride128-sz500k (ns)","43.10 (min 41.58, max 44.38)" - "lat_mem_rd-stride128-sz62k (ns)","7.15 (min 6.43, max 7.42)" - "lat_mmap-1m (us)","69.00 (min 65.00, max 76.00)" + "lat_mem_rd-stride128-sz500k (ns)","42.92 (min 42.10, max 44.04)" + "lat_mem_rd-stride128-sz62k (ns)","7.37 (min 7.35, max 7.39)" + "lat_mmap-1m (us)","61.50 (min 58.00, max 71.00)" "lat_ops-double-add (ns)","4.02" "lat_ops-double-div (ns)","22.13 (min 22.12, max 22.14)" - "lat_ops-double-mul (ns)","4.02 (min 4.02, max 4.03)" + "lat_ops-double-mul (ns)","4.02" "lat_ops-float-add (ns)","4.02" - "lat_ops-float-div (ns)","13.07" + "lat_ops-float-div (ns)","13.08 (min 13.06, max 13.11)" "lat_ops-float-mul (ns)","4.02" "lat_ops-int-add (ns)","1.01" "lat_ops-int-bit (ns)","0.67" - "lat_ops-int-div (ns)","6.03 (min 6.03, max 6.04)" - "lat_ops-int-mod (ns)","6.37 (min 6.37, max 6.39)" - "lat_ops-int-mul (ns)","4.34 (min 4.32, max 4.42)" + "lat_ops-int-div (ns)","6.04 (min 6.03, max 6.05)" + "lat_ops-int-mod (ns)","6.38 (min 6.37, max 6.40)" + "lat_ops-int-mul (ns)","4.34 (min 4.27, max 4.38)" "lat_ops-int64-add (ns)","1.01" "lat_ops-int64-bit (ns)","0.67" - "lat_ops-int64-div (ns)","9.56 (min 9.55, max 9.57)" - "lat_ops-int64-mod (ns)","7.38 (min 7.37, max 7.39)" - "lat_ops-int64-mul (ns)","5.01 (min 4.99, max 5.13)" - "lat_pagefault (us)","1.81 (min 1.76, max 1.84)" - "lat_pipe (us)","26.34 (min 25.73, max 27.06)" - "lat_proc-exec (us)","1176.48 (min 1144.00, max 1221.00)" - "lat_proc-fork (us)","1019.08 (min 1002.83, max 1054.80)" + "lat_ops-int64-div (ns)","9.55" + "lat_ops-int64-mod (ns)","7.38 (min 7.37, max 7.38)" + "lat_ops-int64-mul (ns)","5.05 (min 4.98, max 5.11)" + "lat_pagefault (us)","1.66 (min 1.57, max 1.75)" + "lat_pipe (us)","21.07 (min 20.07, max 21.69)" + "lat_proc-exec (us)","1192.27 (min 1137.20, max 1275.60)" + "lat_proc-fork (us)","1074.68 (min 1031.00, max 1137.60)" "lat_proc-proccall (us)","0.01" - "lat_select (us)","46.72 (min 46.13, max 47.43)" - "lat_sem (us)","3.45 (min 3.16, max 3.71)" - "lat_sig-catch (us)","5.78 (min 5.65, max 5.93)" - "lat_sig-install (us)","0.94 (min 0.93, max 1.00)" - "lat_sig-prot (us)","0.75 (min 0.69, max 0.81)" - "lat_syscall-fstat (us)","2.55 (min 2.49, max 2.67)" - "lat_syscall-null (us)","0.62 (min 0.60, max 0.68)" - "lat_syscall-open (us)","386.20 (min 345.40, max 419.63)" - "lat_syscall-read (us)","0.83 (min 0.82, max 0.88)" - "lat_syscall-stat (us)","6.88 (min 6.48, max 7.35)" - "lat_syscall-write (us)","0.79 (min 0.77, max 0.83)" - "lat_tcp (us)","1.12 (min 1.10, max 1.16)" - "lat_unix (us)","42.81 (min 37.94, max 47.17)" - "latency_for_0.50_mb_block_size (nanosec)","43.10 (min 41.58, max 44.38)" - "latency_for_1.00_mb_block_size (nanosec)","23.69 (min 0.00, max 48.23)" - "pipe_bandwidth (mb\s)","540.02 (min 518.86, max 558.01)" - "pipe_latency (microsec)","26.34 (min 25.73, max 27.06)" + "lat_select (us)","42.85 (min 42.28, max 43.92)" + "lat_sem (us)","2.57 (min 1.71, max 3.19)" + "lat_sig-catch (us)","6.09 (min 5.83, max 6.30)" + "lat_sig-install (us)","0.94 (min 0.87, max 1.10)" + "lat_sig-prot (us)","0.81 (min 0.66, max 0.94)" + "lat_syscall-fstat (us)","2.31 (min 2.13, max 2.49)" + "lat_syscall-null (us)","0.52 (min 0.44, max 0.65)" + "lat_syscall-open (us)","345.64 (min 311.94, max 360.53)" + "lat_syscall-read (us)","0.80 (min 0.69, max 0.89)" + "lat_syscall-stat (us)","5.88 (min 5.61, max 6.20)" + "lat_syscall-write (us)","0.76 (min 0.66, max 0.87)" + "lat_tcp (us)","0.97 (min 0.75, max 1.22)" + "lat_unix (us)","29.27 (min 24.87, max 39.15)" + "latency_for_0.50_mb_block_size (nanosec)","42.92 (min 42.10, max 44.04)" + "latency_for_1.00_mb_block_size (nanosec)","23.50 (min 0.00, max 48.08)" + "pipe_bandwidth (mb\s)","559.24 (min 519.56, max 584.19)" + "pipe_latency (microsec)","21.07 (min 20.07, max 21.69)" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","46.72 (min 46.13, max 47.43)" - "semaphore_latency (microsec)","3.45 (min 3.16, max 3.71)" - "signal_handler_latency (microsec)","0.94 (min 0.93, max 1.00)" - "signal_handler_overhead (microsec)","5.78 (min 5.65, max 5.93)" - "tcp_ip_connection_cost_to_localhost (microsec)","75.76 (min 74.84, max 78.74)" - "tcp_latency_using_localhost (microsec)","1.12 (min 1.10, max 1.16)" + "select_on_200_tcp_fds (microsec)","42.85 (min 42.28, max 43.92)" + "semaphore_latency (microsec)","2.57 (min 1.71, max 3.19)" + "signal_handler_latency (microsec)","0.94 (min 0.87, max 1.10)" + "signal_handler_overhead (microsec)","6.09 (min 5.83, max 6.30)" + "tcp_ip_connection_cost_to_localhost (microsec)","69.65 (min 68.59, max 71.76)" + "tcp_latency_using_localhost (microsec)","0.97 (min 0.75, max 1.22)" Dhrystone -^^^^^^^^^ +========= + Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. For standard ARM cores the DMIPS/MHz score will be identical with the same compiler and flags. @@ -237,29 +194,31 @@ cores the DMIPS/MHz score will be identical with the same compiler and flags. :header: "Benchmarks","am64xx-hsevm: perf" "cpu_clock (mhz)","1000.00" - "dhrystone_per_mhz (dmips/mhz)","2.90" - "dhrystone_per_second (dhrystonep)","5128205.00" + "dhrystone_per_mhz (dmips/mhz)","2.76 (min 2.70, max 2.80)" + "dhrystone_per_second (dhrystonep)","4855981.60 (min 4761905.00, max 5000000.00)" Whetstone -^^^^^^^^^ +========= .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "whetstone (mips)","4583.33 (min 3333.30, max 5000.00)" + "whetstone (mips)","3611.08 (min 3333.30, max 5000.00)" Linpack -^^^^^^^ +======= + Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "linpack (kflops)","407927.00 (min 406745.00, max 408901.00)" + "linpack (kflops)","411621.33 (min 411424.00, max 411819.00)" CoreMarkPro -^^^^^^^^^^^ +=========== + CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, @@ -269,98 +228,52 @@ and floating-point workloads, and data sets for utilizing larger memory subsyste .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "cjpeg-rose7-preset (workloads/)","29.58 (min 29.41, max 29.67)" + "cjpeg-rose7-preset (workloads/)","29.60 (min 29.41, max 29.85)" "core (workloads/)","0.21" - "coremark-pro ()","588.47 (min 585.75, max 593.10)" - "linear_alg-mid-100x100-sp (workloads/)","10.41 (min 10.41, max 10.42)" - "loops-all-mid-10k-sp (workloads/)","0.48" - "nnet_test (workloads/)","0.77 (min 0.76, max 0.77)" - "parser-125k (workloads/)","5.46 (min 5.21, max 5.62)" - "radix2-big-64k (workloads/)","20.60 (min 19.46, max 21.53)" - "sha-test (workloads/)","57.72 (min 57.47, max 57.80)" - "zip-test (workloads/)","15.47 (min 15.38, max 15.63)" + "coremark-pro ()","592.17 (min 585.01, max 602.59)" + "linear_alg-mid-100x100-sp (workloads/)","10.41 (min 10.38, max 10.43)" + "loops-all-mid-10k-sp (workloads/)","0.48 (min 0.48, max 0.49)" + "nnet_test (workloads/)","0.76 (min 0.76, max 0.77)" + "parser-125k (workloads/)","5.61 (min 5.29, max 6.17)" + "radix2-big-64k (workloads/)","20.77 (min 20.27, max 21.14)" + "sha-test (workloads/)","57.80" + "zip-test (workloads/)","15.75 (min 15.63, max 16.13)" .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "cjpeg-rose7-preset (workloads/)","55.33 (min 38.46, max 59.17)" - "core (workloads/)","0.43 (min 0.42, max 0.43)" - "coremark-pro ()","1041.66 (min 1003.86, max 1052.62)" - "linear_alg-mid-100x100-sp (workloads/)","20.81 (min 20.78, max 20.84)" - "loops-all-mid-10k-sp (workloads/)","0.88 (min 0.87, max 0.89)" - "nnet_test (workloads/)","1.54" - "parser-125k (workloads/)","5.86 (min 5.56, max 6.15)" - "radix2-big-64k (workloads/)","32.59 (min 31.39, max 33.66)" - "sha-test (workloads/)","115.13 (min 114.94, max 116.28)" - "zip-test (workloads/)","28.17 (min 27.78, max 28.57)" - -MultiBench -^^^^^^^^^^ -MultiBench™ is a suite of benchmarks that allows processor and system designers to -analyze, test, and improve multicore processors. It uses three forms of concurrency: -Data decomposition: multiple threads cooperating on achieving a unified goal and -demonstrating a processor’s support for fine grain parallelism. -Processing multiple data streams: uses common code running over multiple threads and -demonstrating how well a processor scales over scalable data inputs. -Multiple workload processing: shows the scalability of general-purpose processing, -demonstrating concurrency over both code and data. -MultiBench combines a wide variety of application-specific workloads with the EEMBC -Multi-Instance-Test Harness (MITH), compatible and portable with most any multicore -processors and operating systems. MITH uses a thread-based API (POSIX-compliant) to -establish a common programming model that communicates with the benchmark through an -abstraction layer and provides a flexible interface to allow a wide variety of -thread-enabled workloads to be tested. - -.. csv-table:: Multibench Benchmarks - :header: "Benchmarks","am64xx-hsevm: perf" + "cjpeg-rose7-preset (workloads/)","58.09 (min 57.47, max 58.82)" + "core (workloads/)","0.43" + "coremark-pro ()","1037.34 (min 981.69, max 1065.66)" + "linear_alg-mid-100x100-sp (workloads/)","20.81 (min 20.78, max 20.83)" + "loops-all-mid-10k-sp (workloads/)","0.88" + "nnet_test (workloads/)","1.53" + "parser-125k (workloads/)","5.99 (min 4.91, max 6.99)" + "radix2-big-64k (workloads/)","32.45 (min 31.58, max 33.23)" + "sha-test (workloads/)","114.94" + "zip-test (workloads/)","26.01 (min 14.39, max 28.57)" - "4m-check (workloads/)","282.33 (min 280.68, max 284.84)" - "4m-check-reassembly (workloads/)","61.37 (min 61.12, max 61.50)" - "4m-check-reassembly-tcp (workloads/)","36.64 (min 36.39, max 36.93)" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","14.60 (min 14.57, max 14.63)" - "4m-check-reassembly-tcp-x264w2 (workloads/)","0.75 (min 0.74, max 0.75)" - "4m-cmykw2 (workloads/)","85.93 (min 85.80, max 86.17)" - "4m-cmykw2-rotatew2 (workloads/)","17.54 (min 16.62, max 18.35)" - "4m-reassembly (workloads/)","56.76 (min 56.34, max 57.24)" - "4m-rotatew2 (workloads/)","20.14 (min 18.86, max 21.70)" - "4m-tcp-mixed (workloads/)","84.66 (min 83.77, max 85.56)" - "4m-x264w2 (workloads/)","0.77 (min 0.76, max 0.77)" - "empty-wld (workloads/)","1.00" - "idct-4m (workloads/)","13.67 (min 13.65, max 13.68)" - "idct-4mw1 (workloads/)","13.67 (min 13.65, max 13.68)" - "ippktcheck-4m (workloads/)","282.00 (min 280.96, max 284.51)" - "ippktcheck-4mw1 (workloads/)","282.27 (min 281.37, max 284.61)" - "ipres-4m (workloads/)","72.04 (min 71.70, max 72.39)" - "ipres-4mw1 (workloads/)","72.29 (min 72.05, max 72.50)" - "md5-4m (workloads/)","19.98 (min 19.88, max 20.02)" - "md5-4mw1 (workloads/)","20.01 (min 19.94, max 20.06)" - "rgbcmyk-4m (workloads/)","44.75 (min 44.72, max 44.77)" - "rgbcmyk-4mw1 (workloads/)","44.72 (min 44.69, max 44.77)" - "rotate-4ms1 (workloads/)","16.62 (min 16.49, max 16.78)" - "rotate-4ms1w1 (workloads/)","16.57 (min 16.30, max 16.74)" - "rotate-4ms64 (workloads/)","16.83 (min 16.77, max 16.95)" - "rotate-4ms64w1 (workloads/)","16.82 (min 16.77, max 16.92)" - "x264-4mq (workloads/)","0.41" - "x264-4mqw1 (workloads/)","0.41" +| Boot-time Measurement ---------------------- +===================== Boot media: MMCSD -^^^^^^^^^^^^^^^^^ +----------------- .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am64xx-hsevm: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","22.36 (min 20.60, max 27.42)" + "Linux boot time from SD with default rootfs (20 boot cycles)","28.67 (min 22.45, max 36.73)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. | Ethernet ------------------ -Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html +======== + +Ethernet performance benchmarks were measured using :command:`netperf` 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, @@ -370,11 +283,11 @@ CPU utilization was captured as the total percentage used across all cores on th while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. -In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth +In this scenario, :command:`netperf` options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: -:: +.. code-block:: console burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) burst_size = 500000000 / 8 / 1472 / 100 = 425 @@ -382,21 +295,21 @@ no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B data wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the :command:`netperf` test with no bandwidth limit (remove -b/-w options). -In order to start a netperf client on one device, the other device must have netserver running. -To start netserver: +In order to start a :command:`netperf` client on one device, the other device must have :command:`netserver` running. +To start :command:`netserver`: -:: +.. code-block:: console netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger :command:`netperf` clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test. -:: +.. code-block:: console #!/bin/bash for i in 1 @@ -408,135 +321,133 @@ additional timing measurements during the test. -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for +Running the following commands will trigger :command:`netperf` clients to measure UDP burst performance for 60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run :command:`netperf` client from DUT and start :command:`netserver` on tester. -:: +.. code-block:: console netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run :command:`netperf` client from tester and start :command:`netserver` on DUT. -:: +.. code-block:: console netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE +CPSW/CPSW2g/CPSW3g Ethernet +--------------------------- -CPSW/CPSW2g/CPSW3g Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -- CPSW2g: AM65x, J7200, J721e - CPSW3g: AM64x -.. rubric:: TCP Bidirectional Throughput - :name: CPSW2g-tcp-bidirectional-throughput +TCP Bidirectional Throughput +^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1057.82 (min 989.78, max 1130.34)","97.00 (min 78.90, max 99.90)" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1100.48 (min 1064.93, max 1125.49)","99.80 (min 99.68, max 99.87)" + +TCP Bidirectional Throughput Interrupt Pacing +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +.. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing + :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" + + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1129.68 (min 1104.10, max 1180.26)","92.57 (min 77.95, max 98.21)" + +UDP Throughput +^^^^^^^^^^^^^^ .. csv-table:: CPSW2g UDP Egress Throughput 0 loss - :header: "UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "64","39.82","77.00","89.91" - "128","75.5","74.00","89.57" - "256","148.51","73.00","88.11" - "1024","575.58","70.00","91.17" - "1472","583.88","48.00","84.69" + "64","","31.91 (min 16.92, max 46.99)","62.25 (min 33.00, max 92.00)","71.14 (min 54.01, max 87.52)" + "128","","62.77 (min 33.94, max 92.73)","61.50 (min 33.00, max 91.00)","69.80 (min 52.96, max 86.37)" + "256","","91.10 (min 53.10, max 171.91)","44.50 (min 26.00, max 84.00)","56.55 (min 31.44, max 87.70)" + "1024","","420.23 (min 273.76, max 655.96)","51.00 (min 33.00, max 80.00)","66.26 (min 53.48, max 89.14)" + "1518","","443.96 (min 254.51, max 628.95)","36.50 (min 21.00, max 52.00)","71.19 (min 53.13, max 89.29)" .. csv-table:: CPSW2g UDP Ingress Throughput 0 loss - :header: "UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "64","2.41","5.00","0.14" - "128","4.81","5.00","0.55" - "256","10.85","5.00","0.3" - "1024","43.42","5.00","0.15" - "1472","62.41","5.00","3.28" + "128","","4.89 (min 4.40, max 5.43)","4.50 (min 4.00, max 5.00)","2.05 (min 0.20, max 7.47)" + "256","","10.60 (min 10.24, max 11.06)","5.00","1.86 (min 0.13, max 6.88)" + "1024","","42.05 (min 41.78, max 42.60)","5.00","3.74 (min 0.94, max 7.58)" + "1518","","63.29 (min 62.41, max 64.76)","5.00","7.24 (min 7.04, max 7.31)" -ICSSG Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^ +.. csv-table:: CPSW2g UDP Ingress Throughput possible loss + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)","am64xx-hsevm: Packet Loss %" + + "128","","92.68 (min 84.84, max 97.29)","90.50 (min 83.00, max 95.00)","66.28 (min 64.69, max 67.86)","72.37 (min 68.17, max 79.35)" + "256","","190.43 (min 175.08, max 202.24)","93.00 (min 85.00, max 99.00)","67.43 (min 65.94, max 68.79)","69.29 (min 67.47, max 71.90)" + "1024","","586.35 (min 540.95, max 621.28)","71.67 (min 66.00, max 76.00)","66.59 (min 65.50, max 68.59)","30.88 (min 27.41, max 35.21)" + "1518","","634.14 (min 592.61, max 686.45)","53.75 (min 50.00, max 58.00)","65.50 (min 64.50, max 66.73)","32.49 (min 28.27, max 35.88)" + +ICSSG Ethernet +-------------- -.. rubric:: TCP Bidirectional Throughput - :name: tcp-bidirectional-throughput +TCP Bidirectional Throughput +^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: ICSSG TCP Bidirectional Throughput :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","1007.16 ","99.64" + "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","812.01 (min 355.31, max 1121.74)","83.12 (min 74.78, max 99.72)" -.. rubric:: TCP Bidirectional Throughput Interrupt Pacing - :name: ICSSG-tcp-bidirectional-throughput-interrupt-pacing +TCP Bidirectional Throughput Interrupt Pacing +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: ICSSG TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","1108.21 ","98.21" + "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","528.47 (min 363.92, max 1021.70)","50.15 (min 38.53, max 84.08)" -.. rubric:: UDP Egress Throughput - :name: udp-egress-throughput-0-loss +UDP Egress Throughput +^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: ICSSG UDP Egress Throughput 0 loss - :header: "Frame Size(bytes)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "64","30.77 ","60.13 ","73.74 " - "128","66.04 ","64.67 ","79.53 " - "256","126.02 ","61.50 ","77.46 " - "1024","495.53 ","60.63 ","77.25 " - "1472","789.13 ","67 ","89.64 " + "64","","18.70 (min 16.63, max 19.65)","36.25 (min 32.00, max 38.00)","53.66 (min 53.49, max 53.81)" + "128","","37.99 (min 35.64, max 39.58)","37.25 (min 35.00, max 39.00)","53.92 (min 53.64, max 54.23)" + "256","","122.26 (min 70.86, max 173.65)","59.75 (min 35.00, max 85.00)","71.04 (min 53.56, max 89.45)" + "1024","","421.89 (min 20.48, max 646.48)","51.67 (min 3.00, max 79.00)","59.85 (min 0.95, max 89.47)" + "1472","","225.95 (min 9.42, max 866.11)","19.25 (min 1.00, max 74.00)","23.84 (min 0.19, max 88.88)" -.. rubric:: UDP Ingress Throughput - :name: udp-ingress-throughput-0-loss +UDP Ingress Throughput +^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: ICSSG UDP Ingress Throughput 0 loss - :header: "Frame Size(bytes)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load %" - - "64","2.15 ","4.14 ","1.49 " - "128","4.96 ","4.78 ","3.72 " - "256","10.21 ","5.00 ","2.97 " - "1024","44.54 ","5.25 ","6.07 " - "1472","631.74 ","54 ","68.59 " - -.. rubric:: Switch Mode - :name: icssg-switch-mode - -.. csv-table:: ICSSG Switch Mode Forwarding - :header: "Mode","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (SENDER)","am64xx-hsevm: CPU Load % (FORWARDING)","am64xx-hsevm: CPU Load % (RECIEVER)" + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load %" - "Switch with HW Offload","938","0.3","0.12","37.74" - -.. rubric:: HSR Mode - :name: icssg-hsr-mode - -.. csv-table:: ICSSG HSR Mode Forwarding - :header: "Mode","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (SENDER)","am64xx-hsevm: CPU Load % (FORWARDING)","am64xx-hsevm: CPU Load % (RECIEVER)" - - "HSR with HW Offload","412","47.45","0.13","44.52" - "HSR with SW Offload","387","39.58","27.13","48.74" + "64","","1.55 (min 1.43, max 1.64)","3.00","2.63 (min 0.19, max 6.18)" + "128","","4.43 (min 4.30, max 4.71)","4.25 (min 4.00, max 5.00)","1.16 (min 0.22, max 2.30)" + "256","","10.14 (min 9.63, max 10.65)","5.00","4.54 (min 0.32, max 6.49)" + "1024","","43.22 (min 42.60, max 43.42)","5.00","6.99 (min 5.41, max 9.72)" + "1472","","99.50 (min 61.23, max 181.34)","8.25 (min 5.00, max 15.00)","11.03 (min 6.91, max 20.94)" | -OSPI Flash Driver ------------------ - - -.. rubric:: AM64XX-EVM - :name: am64xx-evm-ospi +OSPI +==== -.. rubric:: RAW - :name: am64xx-evm-ospi-raw +OSPI RAW +-------- .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am64xx-hsevm: Raw Read Throughput (Mbytes/sec)" - "50","139.89 (min 113.64, max 147.06)" + "50","120.03 (min 34.72, max 142.86)" + +| + +EMMC +==== -EMMC Driver ------------ .. warning:: **IMPORTANT**: The performance numbers can be severely affected if the media is @@ -546,51 +457,51 @@ EMMC Driver re-mount in async mode. EMMC EXT4 FIO 1G -^^^^^^^^^^^^^^^^ +---------------- .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "1m","60.75 (min 59.60, max 61.40)","4.00 (min 3.73, max 4.22)","175.00","7.20 (min 6.77, max 7.82)" - "4m","60.83 (min 59.80, max 61.70)","2.88 (min 2.78, max 2.95)","174.75 (min 174.00, max 175.00)","5.76 (min 5.30, max 6.12)" - "4k","49.86 (min 49.10, max 50.40)","52.21 (min 51.71, max 52.53)","56.18 (min 56.00, max 56.40)","47.11 (min 46.00, max 49.21)" - "256k","60.75 (min 59.70, max 61.50)","5.94 (min 5.65, max 6.25)","174.00","8.90 (min 8.58, max 9.48)" + "1m","60.85 (min 59.80, max 61.40)","4.01 (min 3.82, max 4.19)","175.00","7.54 (min 6.84, max 8.58)" + "4m","60.92 (min 60.10, max 61.70)","3.00 (min 2.94, max 3.09)","175.00","6.21 (min 6.08, max 6.67)" + "4k","51.03 (min 50.40, max 51.40)","47.71 (min 47.11, max 48.21)","56.20 (min 55.90, max 56.50)","43.39 (min 42.04, max 44.59)" + "256k","60.88 (min 59.70, max 61.40)","5.90 (min 5.76, max 6.02)","174.00","8.94 (min 8.35, max 9.48)" EMMC EXT4 -""""""""" +--------- .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "102400","53.87 (min 49.65, max 55.64)","11.09 (min 9.38, max 15.25)","176.52 (min 171.28, max 177.31)","29.36 (min 28.30, max 29.63)" - "262144","53.52 (min 49.20, max 55.06)","11.06 (min 9.50, max 15.22)","181.79 (min 181.43, max 182.11)","30.36 (min 29.36, max 31.53)" - "524288","53.64 (min 49.72, max 55.58)","10.94 (min 9.33, max 14.88)","182.88 (min 182.70, max 183.02)","27.74 (min 26.13, max 29.20)" - "1048576","53.83 (min 49.26, max 55.68)","10.72 (min 9.28, max 14.60)","182.91 (min 182.86, max 183.00)","26.93 (min 25.45, max 28.57)" - "5242880","53.93 (min 49.61, max 56.34)","10.61 (min 9.26, max 14.60)","182.80 (min 182.62, max 182.95)","27.64 (min 26.61, max 28.57)" + "102400","53.52 (min 50.76, max 55.23)","10.85 (min 9.02, max 15.05)","177.10 (min 170.77, max 178.65)","29.75 (min 26.36, max 31.19)" + "262144","53.44 (min 49.12, max 55.01)","11.09 (min 9.71, max 14.70)","180.91 (min 177.95, max 182.21)","30.50 (min 27.93, max 33.04)" + "524288","53.65 (min 49.96, max 55.45)","10.94 (min 9.33, max 14.80)","182.81 (min 182.59, max 182.93)","27.26 (min 24.55, max 28.83)" + "1048576","53.88 (min 49.84, max 55.29)","10.68 (min 9.47, max 14.43)","182.82 (min 182.64, max 182.99)","28.13 (min 25.89, max 29.46)" + "5242880","54.16 (min 49.90, max 55.64)","10.63 (min 9.14, max 14.63)","182.84 (min 182.76, max 182.90)","27.89 (min 27.03, max 28.57)" EMMC VFAT -""""""""" +--------- .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am64xx-hsevm: Write VFAT Throughput (Mbytes/sec)","am64xx-hsevm: Write VFAT CPU Load (%)","am64xx-hsevm: Read VFAT Throughput (Mbytes/sec)","am64xx-hsevm: Read VFAT CPU Load (%)" - "102400","49.34 (min 40.10, max 53.27)","13.58 (min 11.41, max 20.45)","166.32 (min 164.99, max 167.10)","30.88 (min 29.20, max 31.90)" - "262144","50.55 (min 41.76, max 53.49)","13.71 (min 11.72, max 20.37)","168.61 (min 161.34, max 171.00)","35.58 (min 29.57, max 50.82)" - "524288","50.67 (min 41.94, max 53.88)","14.01 (min 11.56, max 20.24)","169.98 (min 169.68, max 170.20)","27.60 (min 27.50, max 28.10)" - "1048576","50.86 (min 41.78, max 54.00)","13.53 (min 11.59, max 20.00)","169.78 (min 169.62, max 169.88)","27.34 (min 26.05, max 28.10)" - "5242880","50.75 (min 41.22, max 54.11)","13.44 (min 11.66, max 19.75)","170.06 (min 169.90, max 170.19)","27.70 (min 27.27, max 28.33)" + "102400","48.54 (min 40.00, max 51.49)","12.98 (min 10.62, max 19.53)","168.22 (min 167.35, max 168.71)","31.87 (min 28.81, max 34.48)" + "262144","50.03 (min 40.70, max 53.63)","13.61 (min 11.41, max 19.72)","169.22 (min 167.80, max 170.47)","30.98 (min 27.27, max 37.82)" + "524288","49.42 (min 39.94, max 52.77)","13.40 (min 11.16, max 19.84)","168.08 (min 166.94, max 169.46)","27.37 (min 25.20, max 28.69)" + "1048576","48.91 (min 40.22, max 51.96)","13.10 (min 10.95, max 19.49)","166.64 (min 165.06, max 168.68)","27.83 (min 25.62, max 28.80)" + "5242880","49.84 (min 40.55, max 54.04)","13.15 (min 11.19, max 19.67)","166.79 (min 165.93, max 167.42)","28.51 (min 27.87, max 29.03)" -UBoot EMMC Driver ------------------ +UBoot EMMC +---------- .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am64xx-hsevm: Write Throughput (Kbytes/sec)","am64xx-hsevm: Read Throughput (Kbytes/sec)" - "2000000","59674.24 (min 56594.13, max 61134.33)","168800.07 (min 168041.03, max 169782.38)" - "4000000","60491.03 (min 57387.04, max 61768.14)","172690.99 (min 172463.16, max 173375.66)" + "2000000","59848.32 (min 58724.01, max 61134.33)","169053.08 (min 168907.22, max 169782.38)" + "4000000","60759.37 (min 59308.60, max 61651.93)","172614.84 (min 172463.16, max 172918.21)" MMCSD ------ +===== .. warning:: @@ -601,27 +512,27 @@ MMCSD re-mount in async mode. MMC EXT4 FIO 1G -^^^^^^^^^^^^^^^ +--------------- .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "1m","42.00 (min 41.40, max 42.50)","3.43 (min 3.33, max 3.67)","86.94 (min 86.70, max 87.30)","5.02 (min 4.81, max 5.29)" - "4m","41.90 (min 41.20, max 42.70)","2.59 (min 2.51, max 2.67)","86.33 (min 82.30, max 87.10)","3.73 (min 3.38, max 3.99)" - "4k","2.76 (min 2.70, max 2.87)","7.29 (min 7.00, max 8.06)","12.76 (min 12.70, max 12.80)","14.14 (min 13.37, max 14.99)" - "256k","37.41 (min 36.60, max 38.70)","4.53 (min 4.26, max 4.84)","83.15 (min 82.80, max 83.80)","6.69 (min 6.42, max 7.06)" + "1m","42.42 (min 41.30, max 44.00)","3.46 (min 3.11, max 3.71)","86.77 (min 84.80, max 87.90)","5.25 (min 5.00, max 5.40)" + "4m","41.77 (min 40.30, max 42.70)","2.57 (min 2.51, max 2.68)","85.97 (min 82.60, max 87.30)","3.78 (min 3.35, max 4.18)" + "4k","2.77 (min 2.72, max 2.81)","6.96 (min 6.71, max 7.31)","12.82 (min 12.70, max 12.90)","12.76 (min 12.13, max 13.58)" + "256k","38.15 (min 37.80, max 38.70)","4.64 (min 4.50, max 4.76)","83.47 (min 83.10, max 83.90)","6.51 (min 6.19, max 7.09)" MMC EXT4 -^^^^^^^^ +-------- .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am64xx-hsevm: Write Raw Throughput (Mbytes/sec)","am64xx-hsevm: Write Raw CPU Load (%)","am64xx-hsevm: Read Raw Throughput (Mbytes/sec)","am64xx-hsevm: Read Raw CPU Load (%)" - "102400","29.65 (min 27.48, max 32.92)","6.38 (min 5.22, max 9.52)","38.94 (min 37.34, max 41.04)","7.41 (min 7.07, max 7.77)" - "262144","28.96 (min 27.17, max 31.59)","6.52 (min 5.28, max 9.31)","40.55 (min 39.38, max 41.90)","6.57 (min 6.20, max 7.11)" - "524288","29.58 (min 27.10, max 32.84)","6.12 (min 4.95, max 9.82)","45.12 (min 42.60, max 45.56)","6.75 (min 6.36, max 6.99)" - "1048576","29.85 (min 28.58, max 32.27)","6.13 (min 4.97, max 9.06)","45.39 (min 45.16, max 45.52)","6.71 (min 6.36, max 7.17)" - "5242880","29.78 (min 28.51, max 33.57)","5.85 (min 4.81, max 8.70)","45.22 (min 44.88, max 45.51)","6.85 (min 6.51, max 7.33)" + "102400","28.91 (min 27.03, max 32.81)","5.91 (min 4.58, max 8.99)","39.25 (min 37.68, max 40.92)","7.04 (min 6.25, max 7.95)" + "262144","28.72 (min 27.22, max 32.30)","6.29 (min 4.81, max 8.97)","41.56 (min 39.87, max 42.87)","6.69 (min 6.28, max 7.44)" + "524288","29.68 (min 28.38, max 32.45)","6.11 (min 4.99, max 9.66)","45.38 (min 45.00, max 45.57)","6.76 (min 6.05, max 7.38)" + "1048576","30.06 (min 28.89, max 32.57)","6.01 (min 4.93, max 8.69)","45.34 (min 45.16, max 45.45)","6.75 (min 6.33, max 7.17)" + "5242880","31.13 (min 28.88, max 33.47)","6.14 (min 4.65, max 9.08)","45.36 (min 45.00, max 45.55)","6.87 (min 6.37, max 7.21)" The performance numbers were captured using the following: @@ -630,75 +541,75 @@ The performance numbers were captured using the following: | -CRYPTO Driver -------------- +CRYPTO +====== OpenSSL Performance -^^^^^^^^^^^^^^^^^^^ +------------------- .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am64xx-hsevm: throughput (KBytes/Sec)" - "aes-128-cbc","1024","21201.11 (min 19586.05, max 21629.27)" - "aes-128-cbc","16","305.74 (min 295.07, max 317.97)" - "aes-128-cbc","16384","137300.65 (min 134600.02, max 139203.93)" - "aes-128-cbc","256","5427.83 (min 5317.12, max 5639.77)" - "aes-128-cbc","64","1333.09 (min 1290.43, max 1389.82)" - "aes-128-cbc","8192","98276.35 (min 96264.19, max 99693.91)" - "aes-128-ecb","1024","21623.04 (min 20551.00, max 22215.00)" - "aes-128-ecb","16","312.77 (min 297.66, max 329.37)" - "aes-128-ecb","16384","141950.29 (min 140525.57, max 143638.53)" - "aes-128-ecb","256","5490.12 (min 5154.73, max 5683.37)" - "aes-128-ecb","64","1360.48 (min 1284.46, max 1411.31)" - "aes-128-ecb","8192","101527.21 (min 99691.18, max 102948.86)" - "aes-192-cbc","1024","21028.78 (min 19525.97, max 21523.11)" - "aes-192-cbc","16","311.25 (min 294.13, max 320.98)" - "aes-192-cbc","16384","128866.99 (min 127462.06, max 130378.41)" - "aes-192-cbc","256","5524.78 (min 5259.43, max 5622.10)" - "aes-192-cbc","64","1364.05 (min 1286.27, max 1390.40)" - "aes-192-cbc","8192","94148.27 (min 92752.55, max 95668.91)" - "aes-192-ecb","1024","21608.83 (min 20942.51, max 22065.83)" - "aes-192-ecb","16","314.19 (min 301.85, max 325.51)" - "aes-192-ecb","16384","132728.15 (min 132104.19, max 133327.53)" - "aes-192-ecb","256","5513.00 (min 5258.24, max 5673.05)" - "aes-192-ecb","64","1369.78 (min 1305.26, max 1411.67)" - "aes-192-ecb","8192","97255.42 (min 96520.87, max 98402.30)" - "aes-256-cbc","1024","21003.86 (min 18761.39, max 21494.78)" - "aes-256-cbc","16","314.16 (min 297.62, max 322.65)" - "aes-256-cbc","16384","121110.53 (min 120105.64, max 122344.79)" - "aes-256-cbc","256","5537.48 (min 5443.41, max 5665.71)" - "aes-256-cbc","64","1377.82 (min 1358.29, max 1398.68)" - "aes-256-cbc","8192","90706.26 (min 89451.18, max 91501.91)" - "aes-256-ecb","1024","21289.98 (min 20055.38, max 21793.11)" - "aes-256-ecb","16","308.54 (min 300.03, max 320.75)" - "aes-256-ecb","16384","125571.07 (min 124474.71, max 127385.60)" - "aes-256-ecb","256","5471.54 (min 5277.53, max 5680.98)" - "aes-256-ecb","64","1351.16 (min 1297.41, max 1409.83)" - "aes-256-ecb","8192","93359.79 (min 92383.91, max 94601.22)" - "sha256","1024","28245.16 (min 27951.45, max 28524.20)" - "sha256","16","471.00 (min 466.66, max 475.25)" - "sha256","16384","218273.11 (min 216738.47, max 220250.11)" - "sha256","256","7394.91 (min 7313.49, max 7465.56)" - "sha256","64","1871.55 (min 1853.21, max 1889.11)" - "sha256","8192","149173.25 (min 148398.08, max 149848.06)" - "sha512","1024","18903.94 (min 18712.58, max 19061.76)" - "sha512","16","449.39 (min 444.90, max 451.80)" - "sha512","16384","48433.15 (min 48059.73, max 48562.18)" - "sha512","256","6372.07 (min 6279.94, max 6436.78)" - "sha512","64","1803.87 (min 1779.22, max 1823.55)" - "sha512","8192","43742.21 (min 43578.71, max 43835.39)" + "aes-128-cbc","1024","23012.35" + "aes-128-cbc","16","345.75" + "aes-128-cbc","16384","138062.51" + "aes-128-cbc","256","5988.95" + "aes-128-cbc","64","1506.60" + "aes-128-cbc","8192","100433.92" + "aes-128-ecb","1024","22891.52" + "aes-128-ecb","16","345.72" + "aes-128-ecb","16384","144796.33" + "aes-128-ecb","256","5978.97" + "aes-128-ecb","64","1483.33" + "aes-128-ecb","8192","105411.93" + "aes-192-cbc","1024","22946.13" + "aes-192-cbc","16","322.59" + "aes-192-cbc","16384","130547.71" + "aes-192-cbc","256","5987.33" + "aes-192-cbc","64","1483.73" + "aes-192-cbc","8192","96985.09" + "aes-192-ecb","1024","23038.29" + "aes-192-ecb","16","323.43" + "aes-192-ecb","16384","132852.39" + "aes-192-ecb","256","5995.86" + "aes-192-ecb","64","1401.62" + "aes-192-ecb","8192","100335.62" + "aes-256-cbc","1024","20581.72" + "aes-256-cbc","16","320.95" + "aes-256-cbc","16384","122339.33" + "aes-256-cbc","256","5605.12" + "aes-256-cbc","64","1392.06" + "aes-256-cbc","8192","92940.97" + "aes-256-ecb","1024","21534.04" + "aes-256-ecb","16","326.69" + "aes-256-ecb","16384","127320.06" + "aes-256-ecb","256","5676.46" + "aes-256-ecb","64","1498.28" + "aes-256-ecb","8192","97176.23" + "sha256","1024","24602.62" + "sha256","16","407.77" + "sha256","16384","206176.26" + "sha256","256","6433.96" + "sha256","64","1622.78" + "sha256","8192","137366.19" + "sha512","1024","20061.53" + "sha512","16","393.69" + "sha512","16384","79216.64" + "sha512","256","5904.04" + "sha512","64","1574.76" + "sha512","8192","65959.25" .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am64xx-hsevm: CPU Load" - "aes-128-cbc","43.25 (min 42.00, max 44.00)" - "aes-128-ecb","45.13 (min 43.00, max 46.00)" - "aes-192-cbc","43.75 (min 43.00, max 45.00)" - "aes-192-ecb","44.63 (min 44.00, max 45.00)" - "aes-256-cbc","43.63 (min 43.00, max 44.00)" - "aes-256-ecb","43.88 (min 43.00, max 45.00)" - "sha256","93.00" - "sha512","92.50 (min 90.00, max 93.00)" + "aes-128-cbc","44.00" + "aes-128-ecb","45.00" + "aes-192-cbc","43.00" + "aes-192-ecb","44.00" + "aes-256-cbc","42.00" + "aes-256-ecb","43.00" + "sha256","94.00" + "sha512","94.00" Listed for each algorithm are the code snippets used to run each benchmark test. @@ -707,39 +618,3 @@ Listed for each algorithm are the code snippets used to run each benchmark test. time -v openssl speed -elapsed -evp aes-128-cbc | - -RP Message Inter-Processor Communication (IPC) Latency -------------------------------------------------------- - -RP Message latency Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -RP Message latency is the delay measured from sending a round trip echo message from -a Linux application to a remote processor and back. The following measurements use -a RP message length of 1 byte and 490 bytes for comparison. - -The Linux user space application `rpmsg_char_benchmark `__ captures these latency values. - -Test commands used for running IPC latency tests: - -.. code:: console - - rpmsg_char_benchmark -r 2 -n 100000 -m 1 & chrt -f -p 80 $! - -Latencies reported: - -.. csv-table:: - :header: "Remote Processor","Message Size (in bytes)","Average round trip (usecs)","Max round trip (usecs)" - - "R5F0_0","1","34","165" - "R5F0_0","490","155","279" - "M4F","1","45","138" - "M4F","490","285","408" - -.. image:: img/R5_0_0_m1_ipc_latency.png - -.. image:: img/R5_0_0_m490_ipc_latency.png - -.. image:: img/m4_0_m1_ipc_latency.png - -.. image:: img/m4_0_m490_ipc_latency.png