[InlineCost] Print inline cost for invoke call sites as well #114476
Merged
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
Previously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support.
llvmbot
added
llvm:analysis
Member
|
@llvm/pr-subscribers-llvm-transforms Author: Min-Yih Hsu (mshockwave) ChangesPreviously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support. Full diff: https://github.com/llvm/llvm-project/pull/114476.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/InlineCost.cpp b/llvm/lib/Analysis/InlineCost.cpp
index dd9ac910456ade..22bb406c01a4ed 100644
--- a/llvm/lib/Analysis/InlineCost.cpp
+++ b/llvm/lib/Analysis/InlineCost.cpp
@@ -3257,16 +3257,16 @@ InlineCostAnnotationPrinterPass::run(Function &F,
const InlineParams Params = llvm::getInlineParams();
for (BasicBlock &BB : F) {
for (Instruction &I : BB) {
- if (CallInst *CI = dyn_cast<CallInst>(&I)) {
- Function *CalledFunction = CI->getCalledFunction();
+ if (auto *CB = dyn_cast<CallBase>(&I)) {
+ Function *CalledFunction = CB->getCalledFunction();
if (!CalledFunction || CalledFunction->isDeclaration())
continue;
OptimizationRemarkEmitter ORE(CalledFunction);
- InlineCostCallAnalyzer ICCA(*CalledFunction, *CI, Params, TTI,
+ InlineCostCallAnalyzer ICCA(*CalledFunction, *CB, Params, TTI,
GetAssumptionCache, nullptr, &PSI, &ORE);
ICCA.analyze();
OS << " Analyzing call of " << CalledFunction->getName()
- << "... (caller:" << CI->getCaller()->getName() << ")\n";
+ << "... (caller:" << CB->getCaller()->getName() << ")\n";
ICCA.print(OS);
OS << "\n";
}
diff --git a/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll b/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
index 5c07159822ca1c..f59a565ed34bff 100644
--- a/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
+++ b/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
@@ -33,3 +33,34 @@ define ptr @main() {
%2 = call ptr @foo()
ret ptr %1
}
+
+; Make sure it also analyzes invoke call sites.
+
+; CHECK: Analyzing call of g... (caller:f)
+; CHECK: define i32 @g(i32 %v) {
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: %p = icmp ugt i32 %v, 35
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: %r = select i1 %p, i32 %v, i32 7
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: ret i32 %r
+; CHECK: }
+define i32 @g(i32 %v) {
+ %p = icmp ugt i32 %v, 35
+ %r = select i1 %p, i32 %v, i32 7
+ ret i32 %r
+}
+
+define void @f(i32 %v, ptr %dst) personality ptr @__gxx_personality_v0 {
+ %v1 = invoke i32 @g(i32 %v)
+ to label %bb1 unwind label %bb2
+bb1:
+ store i32 %v1, ptr %dst
+ ret void
+bb2:
+ %lpad.loopexit80 = landingpad { ptr, i32 }
+ cleanup
+ ret void
+}
+
+declare i32 @__gxx_personality_v0(...)
|
Member
|
@llvm/pr-subscribers-llvm-analysis Author: Min-Yih Hsu (mshockwave) ChangesPreviously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support. Full diff: https://github.com/llvm/llvm-project/pull/114476.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/InlineCost.cpp b/llvm/lib/Analysis/InlineCost.cpp
index dd9ac910456ade..22bb406c01a4ed 100644
--- a/llvm/lib/Analysis/InlineCost.cpp
+++ b/llvm/lib/Analysis/InlineCost.cpp
@@ -3257,16 +3257,16 @@ InlineCostAnnotationPrinterPass::run(Function &F,
const InlineParams Params = llvm::getInlineParams();
for (BasicBlock &BB : F) {
for (Instruction &I : BB) {
- if (CallInst *CI = dyn_cast<CallInst>(&I)) {
- Function *CalledFunction = CI->getCalledFunction();
+ if (auto *CB = dyn_cast<CallBase>(&I)) {
+ Function *CalledFunction = CB->getCalledFunction();
if (!CalledFunction || CalledFunction->isDeclaration())
continue;
OptimizationRemarkEmitter ORE(CalledFunction);
- InlineCostCallAnalyzer ICCA(*CalledFunction, *CI, Params, TTI,
+ InlineCostCallAnalyzer ICCA(*CalledFunction, *CB, Params, TTI,
GetAssumptionCache, nullptr, &PSI, &ORE);
ICCA.analyze();
OS << " Analyzing call of " << CalledFunction->getName()
- << "... (caller:" << CI->getCaller()->getName() << ")\n";
+ << "... (caller:" << CB->getCaller()->getName() << ")\n";
ICCA.print(OS);
OS << "\n";
}
diff --git a/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll b/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
index 5c07159822ca1c..f59a565ed34bff 100644
--- a/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
+++ b/llvm/test/Transforms/Inline/inline-cost-annotation-pass.ll
@@ -33,3 +33,34 @@ define ptr @main() {
%2 = call ptr @foo()
ret ptr %1
}
+
+; Make sure it also analyzes invoke call sites.
+
+; CHECK: Analyzing call of g... (caller:f)
+; CHECK: define i32 @g(i32 %v) {
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: %p = icmp ugt i32 %v, 35
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: %r = select i1 %p, i32 %v, i32 7
+; CHECK: ; cost before = {{.*}}, cost after = {{.*}}, threshold before = {{.*}}, threshold after = {{.*}}, cost delta = {{.*}}
+; CHECK: ret i32 %r
+; CHECK: }
+define i32 @g(i32 %v) {
+ %p = icmp ugt i32 %v, 35
+ %r = select i1 %p, i32 %v, i32 7
+ ret i32 %r
+}
+
+define void @f(i32 %v, ptr %dst) personality ptr @__gxx_personality_v0 {
+ %v1 = invoke i32 @g(i32 %v)
+ to label %bb1 unwind label %bb2
+bb1:
+ store i32 %v1, ptr %dst
+ ret void
+bb2:
+ %lpad.loopexit80 = landingpad { ptr, i32 }
+ cleanup
+ ret void
+}
+
+declare i32 @__gxx_personality_v0(...)
|
aeubanks
approved these changes
Nov 1, 2024
smallp-o-p
pushed a commit
to smallp-o-p/llvm-project
that referenced
this pull request
Nov 3, 2024
…4476) Previously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support.
NoumanAmir657
pushed a commit
to NoumanAmir657/llvm-project
that referenced
this pull request
Nov 4, 2024
…4476) Previously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Previously InlineCostAnnotationPrinter only prints inline cost for call instructions. I don't think there is any reason not to analyze invoke and its callee, and this patch adds such support.