[all-commits] [llvm/llvm-project] 04ebd1: [MC][NFC] Statically allocate storage for decoded ...

[Amir Ayupov via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 04ebd1907c0561831e4fcf2658e1f3614f8cdd77
      https://github.com/llvm/llvm-project/commit/04ebd1907c0561831e4fcf2658e1f3614f8cdd77
  Author: Amir Ayupov <aaupov at fb.com>
  Date:   2024-08-26 (Mon, 26 Aug 2024)

  Changed paths:
    M bolt/lib/Rewrite/PseudoProbeRewriter.cpp
    M llvm/include/llvm/MC/MCPseudoProbe.h
    M llvm/lib/MC/MCPseudoProbe.cpp
    M llvm/tools/llvm-profgen/ProfileGenerator.cpp
    M llvm/tools/llvm-profgen/ProfiledBinary.cpp

  Log Message:
  -----------
  [MC][NFC] Statically allocate storage for decoded pseudo probes and function records

Use #102774 to allocate storage for decoded probes (`PseudoProbeVec`)
and function records (`InlineTreeVec`).

Leverage that to also shrink sizes of `MCDecodedPseudoProbe`:
- Drop Guid since it's accessible via `InlineTree`.

`MCDecodedPseudoProbeInlineTree`:
- Keep track of probes and inlinees using `ArrayRef`s now that probes
  and function records belonging to the same function are allocated
  contiguously.

This reduces peak RSS from 13.7 GiB to 9.7 GiB and pseudo probe parsing
time (as part of perf2bolt) from 15.3s to 9.6s for a large binary with
400MiB .pseudo_probe section containing 43M probes and 25M function
records.

Depends on:
#102774
#102787
#102788

Reviewers: maksfb, rafaelauler, dcci, ayermolo, wlei-llvm

Reviewed By: wlei-llvm

Pull Request: https://github.com/llvm/llvm-project/pull/102789



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications