[llvm] [PredicateInfo] Cache ssa.copy declarations (NFC) (PR #145020)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-llvm-transforms

Author: Nikita Popov (nikic)

<details>
<summary>Changes</summary>

This pass creates a lot of ssa.copy intrinsics, typically for a small set of types. Determining the function type, performing intrinsic name mangling and looking up the declaration has noticeable overhead in this case.

Improve this by caching the declarations by type. I've made this a separate map from CreatedDeclarations, which only tracks the declarations that were newly inserted (but not pre-existing ones).

---
Full diff: https://github.com/llvm/llvm-project/pull/145020.diff


2 Files Affected:

- (modified) llvm/include/llvm/Transforms/Utils/PredicateInfo.h (+2) 
- (modified) llvm/lib/Transforms/Utils/PredicateInfo.cpp (+15-12) 


``````````diff
diff --git a/llvm/include/llvm/Transforms/Utils/PredicateInfo.h b/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
index 670bfaa5ad6fe..1619fa31fb6f4 100644
--- a/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
+++ b/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
@@ -206,6 +206,8 @@ class PredicateInfo {
   DenseMap<const Value *, const PredicateBase *> PredicateMap;
   // The set of ssa_copy declarations we created with our custom mangling.
   SmallSet<AssertingVH<Function>, 20> CreatedDeclarations;
+  // Cache of ssa.copy declaration for a given type.
+  SmallDenseMap<Type *, Function *> DeclarationCache;
 };
 
 /// Printer pass for \c PredicateInfo.
diff --git a/llvm/lib/Transforms/Utils/PredicateInfo.cpp b/llvm/lib/Transforms/Utils/PredicateInfo.cpp
index 97f13e3b26746..074b87265e375 100644
--- a/llvm/lib/Transforms/Utils/PredicateInfo.cpp
+++ b/llvm/lib/Transforms/Utils/PredicateInfo.cpp
@@ -499,6 +499,19 @@ Value *PredicateInfoBuilder::materializeStack(unsigned int &Counter,
     ValInfo->RenamedOp = (RenameStack.end() - Start) == RenameStack.begin()
                              ? OrigOp
                              : (RenameStack.end() - Start - 1)->Def;
+    auto CreateSSACopy = [this](IRBuilderBase &B, Value *Op,
+                                const Twine &Name = "") {
+      auto It = PI.DeclarationCache.try_emplace(Op->getType());
+      if (It.second) {
+        auto NumDecls = F.getParent()->getNumNamedValues();
+        Function *IF = Intrinsic::getOrInsertDeclaration(
+            F.getParent(), Intrinsic::ssa_copy, Op->getType());
+        if (NumDecls != F.getParent()->getNumNamedValues())
+          PI.CreatedDeclarations.insert(IF);
+        It.first->second = IF;
+      }
+      return B.CreateCall(It.first->second, Op, Name);
+    };
     // For edge predicates, we can just place the operand in the block before
     // the terminator. For assume, we have to place it right after the assume
     // to ensure we dominate all uses except assume itself. Always insert
@@ -511,13 +524,8 @@ Value *PredicateInfoBuilder::materializeStack(unsigned int &Counter,
     // represents an invalid module.
     if (isa<PredicateWithEdge>(ValInfo)) {
       IRBuilder<> B(getBranchTerminator(ValInfo));
-      auto NumDecls = F.getParent()->getNumNamedValues();
-      Function *IF = Intrinsic::getOrInsertDeclaration(
-          F.getParent(), Intrinsic::ssa_copy, Op->getType());
-      if (NumDecls != F.getParent()->getNumNamedValues())
-        PI.CreatedDeclarations.insert(IF);
       CallInst *PIC =
-          B.CreateCall(IF, Op, Op->getName() + "." + Twine(Counter++));
+          CreateSSACopy(B, Op, Op->getName() + "." + Twine(Counter++));
       PI.PredicateMap.insert({PIC, ValInfo});
       Result.Def = PIC;
     } else {
@@ -527,12 +535,7 @@ Value *PredicateInfoBuilder::materializeStack(unsigned int &Counter,
       // Insert the predicate directly after the assume. While it also holds
       // directly before it, assume(i1 true) is not a useful fact.
       IRBuilder<> B(PAssume->AssumeInst->getNextNode());
-      auto NumDecls = F.getParent()->getNumNamedValues();
-      Function *IF = Intrinsic::getOrInsertDeclaration(
-          F.getParent(), Intrinsic::ssa_copy, Op->getType());
-      if (NumDecls != F.getParent()->getNumNamedValues())
-        PI.CreatedDeclarations.insert(IF);
-      CallInst *PIC = B.CreateCall(IF, Op);
+      CallInst *PIC = CreateSSACopy(B, Op);
       PI.PredicateMap.insert({PIC, ValInfo});
       Result.Def = PIC;
     }

``````````

</details>


https://github.com/llvm/llvm-project/pull/145020