News for hackers. Stuff that matters.

A comment on SciPy 1.0: fundamental algorithms for scientific computing in Python. Parent comment

Re:SciPy 1.0: fundamental algorithms for scientific computin... by tda on Tuesday February 11, 2020 @07:37AM (#22297962)
Why is (1025, 1025, 1025) so much faster than (1024, 1024, 1024)?
[ Parent | on HN ]
Re:SciPy 1.0: fundamental algorithms for scientific computin... by kmbriedis on Wednesday February 12, 2020 @02:23AM (#22306670)
My guess that it's happening mostly due cache conflicts. With 1024 for a simplified L1 with 32kb you can fit exactly 8 lines of the inner dimmension in the cache, which means that (0,8,0) would have the same cache location as (0, 0, 0), which is bad for tiling
[ Parent | on HN ]