Bernardo Covas
Bernardo Covas
Hi @mmacklin , thanks for the reply Yes! I would have to create it somehow beforehand
Hello @christophercrouzet , @mmacklin Thank you very much!
Hi @mmacklin I see, thank you for the response!
Hello @shi-eric , @mmacklin Somehow, and quite embarisingly so, since posting this, I could not reproduce the issue. It used to happen when some warp processing was done before the...
Also, @shi-eric , this is using standard python 3.10
Hi @shi-eric , Thank you for the help! So maybe I am not understanding the current approach. If a static variable is used, it means that warp never releases the...
Yes that is likely it. I do use numpy quite a bit, and I do also release the gil in my own codebase basically everywhere, so I use the thread...
After printing the thread index inside the kernel in the beginning and end, I can confirm no two kernels are executing simultaneously, as one allways starts and ends before any...
Hi @shi-eric, @mmacklin I see that this has been addressed! Thank you!