[PATCH] D156278: [BOLT] Fine-tuning hash computation for stale matching

[Amir Ayupov via Phabricator via llvm-commits]

Amir added a comment.

Please fix a couple of nits, LGTM otherwise.



================
Comment at: bolt/lib/Core/HashUtilities.cpp:160
+    if (IsX86)
+      Opcode = BC.MIB->getShortOpcode(Opcode);
+
----------------
As we discussed 


================
Comment at: bolt/lib/Target/X86/X86MCPlusBuilder.cpp:2875
+  unsigned getShortOpcode(unsigned Opcode) const override {
+    return X86::getOpcodeForShortImmediateForm(Opcode);
+  }
----------------
This function doesn't do what we wanted it to do (return the short version


================
Comment at: bolt/test/X86/reader-stale-yaml.test:41
+CHECK1:    .LFT[[#BB1:]] (2 instructions, align : 1)
+# Check the overal inference stats.
+CHECK1:  2 out of 7 functions in the binary (28.6%) have non-empty execution profile
----------------



================
Comment at: bolt/test/X86/reader-stale-yaml.test:49
+# verifies that the inference is able to match two blocks (bid=1 and bid=13)
+# using "loose" hashes and then correctly propogate the counts.
 
----------------



Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D156278/new/

https://reviews.llvm.org/D156278