[all-commits] [llvm/llvm-project] 66ba71: [llvm-profdata] Refactoring Sample Profile Reader ...

[William Junda Huang via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 66ba71d913df7f7cd75e92c0c4265932b7c93292
      https://github.com/llvm/llvm-project/commit/66ba71d913df7f7cd75e92c0c4265932b7c93292
  Author: William Huang <williamjhuang at google.com>
  Date:   2023-07-27 (Thu, 27 Jul 2023)

  Changed paths:
    M llvm/include/llvm/ProfileData/SampleProf.h
    M llvm/include/llvm/ProfileData/SampleProfReader.h
    M llvm/lib/ProfileData/ProfileSummaryBuilder.cpp
    M llvm/lib/ProfileData/SampleProf.cpp
    M llvm/lib/ProfileData/SampleProfReader.cpp
    M llvm/lib/ProfileData/SampleProfWriter.cpp
    M llvm/lib/Transforms/IPO/SampleContextTracker.cpp
    M llvm/test/tools/llvm-profdata/Inputs/sample-nametable-after-samples.profdata
    M llvm/test/tools/llvm-profdata/sample-nametable.test
    M llvm/tools/llvm-profdata/llvm-profdata.cpp
    M llvm/tools/llvm-profgen/ProfileGenerator.cpp
    M llvm/unittests/tools/llvm-profdata/CMakeLists.txt
    A llvm/unittests/tools/llvm-profdata/MD5CollisionTest.cpp
    M llvm/unittests/tools/llvm-profdata/OutputSizeLimitTest.cpp

  Log Message:
  -----------
  [llvm-profdata] Refactoring Sample Profile Reader to increase FDO build speed using MD5 as key to Sample Profile map

This is phase 1 of multiple planned improvements on the sample profile loader.   The major change is to use MD5 hash code ((instead of the function itself) as the key to look up the function offset table and the profiles, which significantly reduce the time it takes to construct the map.

The optimization is based on the fact that many practical sample profiles are using MD5 values for function names to reduce profile size, so we shouldn't need to convert the MD5 to a string and then to a SampleContext and use it as the map's key, because it's extremely slow.

Several changes to note:

(1) For non-CS SampleContext, if it is already MD5 string, the hash value will be its integral value, instead of hashing the MD5 again. In phase 2 this is going to be optimized further using a union to represent MD5 function (without converting it to string) and regular function names.

(2) The SampleProfileMap is a wrapper to *map<uint64_t, FunctionSamples>, while providing interface allowing using SampleContext as key, so that existing code still work. It will check for MD5 collision (unlikely but not too unlikely, since we only takes the lower 64 bits) and handle it to at least guarantee compilation correctness (conflicting old profile is dropped, instead of returning an old profile with inconsistent context). Other code should not try to use MD5 as key to access the map directly, because it will not be able to handle MD5 collision at all. (see exception at (5) )

(3) Any SampleProfileMap::emplace() followed by SampleContext assignment if newly inserted, should be replaced with SampleProfileMap::Create(), which does the same thing.

(4) Previously we ensure an invariant that in SampleProfileMap, the key is equal to the Context of the value, for profile map that is eventually being used for output (as in llvm-profdata/llvm-profgen). Since the key became MD5 hash, only the value keeps the context now, in several places where an intermediate SampleProfileMap is created, each new FunctionSample's context is set immediately after insertion, which is necessary to "remember" the context otherwise irretrievable.

(5) When reading a profile, we cache the MD5 values of all functions, because they are used at least twice (one to index into FuncOffsetTable, the other into SampleProfileMap, more if there are additional sections), in this case the SampleProfileMap is directly accessed with MD5 value so that we don't recalculate it each time (expensive)

Performance impact:
When reading a ~1GB extbinary profile (fixed length MD5, not compressed) with 10 million function names and 2.5 million top level functions (non CS functions, each function has varying nesting level from 0 to 20), this patch improves the function offset table loading time by 20%, and improves full profile read by 5%.

Reviewed By: davidxl, snehasish

Differential Revision: https://reviews.llvm.org/D147740