[VPlan] Remove no longer needed VP intrinsic handling in VPWidenIntrinsicRecipe::computeCost. NFCI #137573
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…nsicRecipe::computeCost. NFCI Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
|
@llvm/pr-subscribers-vectorizers @llvm/pr-subscribers-llvm-transforms Author: Luke Lau (lukel97) ChangesWhenever calls were transformed to VP intrinsics with EVL tail folding in #110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in #115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in #110412 and checking that no assertions were still hit. Full diff: https://github.com/llvm/llvm-project/pull/137573.diff 1 Files Affected:
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index 75d056026025a..54990c7c806c4 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -1333,15 +1333,6 @@ InstructionCost VPWidenIntrinsicRecipe::computeCost(ElementCount VF,
for (const auto &[Idx, Op] : enumerate(operands())) {
auto *V = Op->getUnderlyingValue();
if (!V) {
- // Push all the VP Intrinsic's ops into the Argments even if is nullptr.
- // Some VP Intrinsic's cost will assert the number of parameters.
- // Mainly appears in the following two scenarios:
- // 1. EVL Op is nullptr
- // 2. The Argmunt of the VP Intrinsic is also the VP Intrinsic
- if (VPIntrinsic::isVPIntrinsic(VectorIntrinsicID)) {
- Arguments.push_back(V);
- continue;
- }
if (auto *UI = dyn_cast_or_null<CallBase>(getUnderlyingValue())) {
Arguments.push_back(UI->getArgOperand(Idx));
continue;
|
Mel-Chen
approved these changes
Apr 29, 2025
Yes I think so. I'm not sure why this isn't an issue at the moment. But we should be checking |
fhahn
approved these changes
Apr 29, 2025
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/162/builds/21259 Here is the relevant piece of the build log for the reference |
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…nsicRecipe::computeCost. NFCI (llvm#137573) Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…nsicRecipe::computeCost. NFCI (llvm#137573) Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…nsicRecipe::computeCost. NFCI (llvm#137573) Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
GeorgeARM
pushed a commit
to GeorgeARM/llvm-project
that referenced
this pull request
May 7, 2025
…nsicRecipe::computeCost. NFCI (llvm#137573) Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
Ankur-0429
pushed a commit
to Ankur-0429/llvm-project
that referenced
this pull request
May 9, 2025
…nsicRecipe::computeCost. NFCI (llvm#137573) Whenever calls were transformed to VP intrinsics with EVL tail folding in llvm#110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs(). However it turned out that the actual arguments were never used and this assertion was removed in llvm#115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead. The type based cost and value based cost are the same for these VP intrinsics. This was tested by adding back in the transformation code in llvm#110412 and checking that no assertions were still hit.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Whenever calls were transformed to VP intrinsics with EVL tail folding in #110412, this workaround was added in computeCost to avoid an assertion when checking ICA.getArgs().
However it turned out that the actual arguments were never used and this assertion was removed in #115983 afterwards, so it's now fine to leave the arguments empty and use the type based cost instead.
The type based cost and value based cost are the same for these VP intrinsics.
This was tested by adding back in the transformation code in #110412 and checking that no assertions were still hit.