github/ROCm
1
1
Fork 0
mirror of https://github.com/ROCm/ROCm.git synced 2026-04-05 03:01:17 -04:00
Code Issues Packages Projects Releases Wiki Activity

[TEST] Fixed and re-enabled reduce test (#1644)

Browse Source 
Re-enabled reduce test after fixing the %cst stride in the ttgir, and
modifying the sweep parameters to make sure the shape per CTA to be less
than or equal to the tensor shape.
This commit is contained in:
Zahi Moudallal
2023-05-10 15:15:11 -07:00
committed by GitHub
parent 147ec4384d
commit fb40bf1954
4 changed files with 78 additions and 63 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
lib/Analysis/Utility.cpp
View File

@@ -16,12 +16,25 @@ bool ReduceOpHelper::isFastReduction() {
unsigned ReduceOpHelper::getInterWarpSize() {
auto srcReduceDimSize = static_cast<unsigned>(srcShape[axis]);
unsigned sizeIntraWarps = getIntraWarpSize();
return std::min(srcReduceDimSize / sizeIntraWarps,
triton::gpu::getWarpsPerCTA(getSrcLayout())[axis]);
}
unsigned ReduceOpHelper::getIntraWarpSize() {
auto srcReduceDimSize = static_cast<unsigned>(srcShape[axis]);
return std::min(srcReduceDimSize,
triton::gpu::getThreadsPerWarp(getSrcLayout())[axis]);
}
unsigned ReduceOpHelper::getInterWarpSizeWithUniqueData() {
auto srcReduceDimSize = static_cast<unsigned>(srcShape[axis]);
unsigned sizeIntraWarps = getIntraWarpSizeWithUniqueData();
return std::min(srcReduceDimSize / sizeIntraWarps,
triton::gpu::getWarpsPerCTAWithUniqueData(
getSrcLayout(), getSrcShape())[axis]);
}
unsigned ReduceOpHelper::getIntraWarpSize() {
unsigned ReduceOpHelper::getIntraWarpSizeWithUniqueData() {
auto srcReduceDimSize = static_cast<unsigned>(srcShape[axis]);
return std::min(srcReduceDimSize,
triton::gpu::getThreadsPerWarpWithUniqueData(