github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-04-07 03:00:26 -04:00
Code Issues Packages Projects Releases Wiki Activity

don't upcast already upcasted

Browse Source
This commit is contained in:
George Hotz
2023-01-28 14:58:28 -08:00
parent 2f194aadad
commit b134a4f3d1
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tinygrad/llops/ops_gpu.py
View File

@@ -211,7 +211,8 @@ class CLASTKernel(ASTKernel):
# if last dim <= 3 and it's a reduce dim, upcast (loop unrolling)
end_dimension = max([st.shape[-1] for st in self.sts])
if self.first_reduce < self.shape_len and end_dimension > 1 and end_dimension <= 3: self.upcast()
if self.first_reduce < self.shape_len and end_dimension > 1 and end_dimension <= 3 and max([x.size() for i,x in enumerate(self.buftokens) if self.bufs[i] in self.earlybufs]) <= 4:
self.upcast()
# STOP WASTING TIME WITH DOING THE RESHAPES AND PERMUTES BY HAND. KERNEL SEARCH IS THE ONLY WAY IT WILL EVER BE GOOD
# group_for_reduce will have to be better first