Skip to content

Faster true count using AVX2 and AVX512 instructions #6931

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
robert3005 wants to merge 3 commits into
base: develop
Choose a base branch
from

trythis

d3c062d
Select commit
Loading
Failed to load commit list.
Open

Faster true count using AVX2 and AVX512 instructions #6931

trythis
d3c062d
Select commit
Loading
Failed to load commit list.
CodSpeed HQ / CodSpeed Performance Analysis succeeded Mar 14, 2026 in 0s

Performance Gate Passed

⚡ 8 improved benchmarks
✅ 1001 untouched benchmarks
⏩ 1515 skipped benchmarks1

Performance Changes

Mode Benchmark BASE HEAD Efficiency
⚡ Simulation null_count_run_end[(100000, 1024, 0.1)] 65.6 µs 57.8 µs +13.37%
⚡ Simulation null_count_run_end[(100000, 1024, 0.5)] 65.9 µs 58.2 µs +13.31%
⚡ Simulation null_count_run_end[(100000, 256, 0.5)] 75.2 µs 67.4 µs +11.59%
⚡ Simulation null_count_run_end[(100000, 256, 0.01)] 72.8 µs 64.9 µs +12.14%
⚡ Simulation null_count_run_end[(100000, 256, 0.1)] 73.1 µs 65.3 µs +12.06%
⚡ Simulation true_count_vortex_buffer[16384] 3.6 µs 2.4 µs +52.36%
⚡ Simulation true_count_vortex_buffer[65536] 11.8 µs 6.6 µs +79.68%
⚡ Simulation true_count_vortex_buffer[128] 1,013.9 ns 538.6 ns +88.24%

Comparing rk/truecount (d3c062d) with develop (fc4d111)

Open in CodSpeed

Footnotes

  1. 1515 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

View more details on CodSpeed HQ