[llvm-profgen] On-demand symbolization
authorwlei <wlei@fb.com>
Sat, 25 Sep 2021 00:06:39 +0000 (17:06 -0700)
committerwlei <wlei@fb.com>
Tue, 28 Sep 2021 16:09:25 +0000 (09:09 -0700)
commit091c16f76ba1e6341afd717445323f8396b7772f
tree9e9db0365d2d5a26c432aada5582e5010c2244d8
parent70391b3468b8a4a07b49df88d7fa88c9644cda77
[llvm-profgen] On-demand symbolization

Previously we do symbolization for all the functions and actually we only need the symbols that's hit by the samples.

This can significantly speed up the time for large size binary.

Optimization for per-inliner will come along with next patch.

Reviewed By: hoy, wenlei

Differential Revision: https://reviews.llvm.org/D110465
llvm/tools/llvm-profgen/PerfReader.h
llvm/tools/llvm-profgen/ProfiledBinary.cpp
llvm/tools/llvm-profgen/ProfiledBinary.h