Hi. I use concurrent.futures to handle multi-threading/multi-processing. I can get a basic example to work fine with numpy_quaddtype. But when I try my full normal test (that works with standard numpy) I get:
File "/Users/kaplan/miniforge3/envs/pintdev_arm/lib/python3.14/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/kaplan/miniforge3/envs/pintdev_arm/lib/python3.14/concurrent/futures/process.py", line 242, in _process_worker
call_item = call_queue.get(block=True)
File "/Users/kaplan/miniforge3/envs/pintdev_arm/lib/python3.14/multiprocessing/queues.py", line 120, in get
return _ForkingPickler.loads(res)
~~~~~~~~~~~~~~~~~~~~~^^^^^
RuntimeError: Using legacy SETITEM with NULL array object is only supported for basic NumPy DTypes.
I don't know if I am not triggering this function with my basic example or what. I understand that it may be using pickling to pass between the different processes, but don't know if this can be fixed through a change on my end or if this has to be done here or in multiprocessing. Any ideas?
(sorry about the poorly formatted question)
Hi. I use
concurrent.futuresto handle multi-threading/multi-processing. I can get a basic example to work fine withnumpy_quaddtype. But when I try my full normal test (that works with standard numpy) I get:I don't know if I am not triggering this function with my basic example or what. I understand that it may be using pickling to pass between the different processes, but don't know if this can be fixed through a change on my end or if this has to be done here or in
multiprocessing. Any ideas?(sorry about the poorly formatted question)