[all-commits] [llvm/llvm-project] 21a77e: [IR] Reorder Value fields to put the SubclassID fi...

[Reid Kleckner via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 21a77e8a92a8d97f0dfd5f9f4faa7b6bc82887aa
      https://github.com/llvm/llvm-project/commit/21a77e8a92a8d97f0dfd5f9f4faa7b6bc82887aa
  Author: Reid Kleckner <rnk at google.com>
  Date:   2024-01-12 (Fri, 12 Jan 2024)

  Changed paths:
    M llvm/include/llvm/IR/Value.h
    M llvm/lib/IR/Value.cpp

  Log Message:
  -----------
  [IR] Reorder Value fields to put the SubclassID first (#53520)

Placing the class id at offset 0 should make `isa` and `dyn_cast` faster
by eliminating the field offset (previously 0x10) from the memory
operand, saving encoding space on x86, and, in theory, an add micro-op.
You can see the load encodes one byte smaller here:
https://godbolt.org/z/Whvz4can9

The compile time tracker shows some modestly positive results in the
on the `cycle` metric and in the final clang binary size metric:
https://llvm-compile-time-tracker.com/compare.php?from=33b54f01fe32030ff60d661a7a951e33360f82ee&to=2530347a57401744293c54f92f9781fbdae3d8c2&stat=cycles
Clicking through to the per-library size breakdown shows that
instcombine size reduces by 0.68%, which is meaningful, and I believe
instcombine is known to be a hotspot.

It is, however, potentially noise. I still think we should do this,
because notionally, the class id really acts as the vptr of the Value,
and conventionally the vptr is always at offset 0.