[all-commits] [llvm/llvm-project] 6951b7: [clang] Speedup line offset mapping computation

[serge-sans-paille via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 6951b72334bbe4c189c71751edc1e361d7b5632c
      https://github.com/llvm/llvm-project/commit/6951b72334bbe4c189c71751edc1e361d7b5632c
  Author: serge-sans-paille <sguelton at redhat.com>
  Date:   2021-04-07 (Wed, 07 Apr 2021)

  Changed paths:
    M clang/lib/Basic/SourceManager.cpp
    M llvm/test/tools/llvm-objdump/X86/source-interleave-prefix.test

  Log Message:
  -----------
  [clang] Speedup line offset mapping computation

Clang spends a decent amount of time in the LineOffsetMapping::get(...)
function. This function used to be vectorized (through SSE2) then the
optimization got dropped because the sequential version was on-par performance
wise.

This provides an optimization of the sequential version that works on a word at
a time, using (documented) bithacks to provide a portable vectorization.

When preprocessing the sqlite amalgamation, this yields a sweet 3% speedup.

Differential Revision: https://reviews.llvm.org/D99409