[MemProf][NFC] Hoist size computation out of the loop for v3

[snehasish]

Similar to the suggestion in #137394. In this case apply it to the current binary format (v3).

[snehasish]

• [MemProf][NFC] Hoist size computation out of the loop for v3 #137479 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-pgo

Author: Snehasish Kumar (snehasish)

Changes

Similar to the suggestion in #137394. In this case apply it to the current binary format (v3).

---

Full diff: https://github.com/llvm/llvm-project/pull/137479.diff

1 Files Affected:

• (modified) llvm/lib/ProfileData/MemProf.cpp (+2-1)

diff --git a/llvm/lib/ProfileData/MemProf.cpp b/llvm/lib/ProfileData/MemProf.cpp
index 0af08ca51481f..1e956250e5b84 100644
--- a/llvm/lib/ProfileData/MemProf.cpp
+++ b/llvm/lib/ProfileData/MemProf.cpp
@@ -190,12 +190,13 @@ static IndexedMemProfRecord deserializeV3(const MemProfSchema &Schema,
   const uint64_t NumNodes =
       endian::readNext<uint64_t, llvm::endianness::little>(Ptr);
   Record.AllocSites.reserve(NumNodes);
+  const size_t SerializedSize = PortableMemInfoBlock::serializedSize(Schema);
   for (uint64_t I = 0; I < NumNodes; I++) {
     IndexedAllocationInfo Node;
     Node.CSId =
         endian::readNext<LinearCallStackId, llvm::endianness::little>(Ptr);
     Node.Info.deserialize(Schema, Ptr);
-    Ptr += PortableMemInfoBlock::serializedSize(Schema);
+    Ptr += SerializedSize;
     Record.AllocSites.push_back(Node);
   }
 

[mingmingl-llvm]

thanks for the change!

[snehasish]

Merge activity

• Apr 28, 2:14 PM EDT: A user started a stack merge that includes this pull request via Graphite.
• Apr 28, 2:16 PM EDT: @snehasish merged this pull request with Graphite.