[all-commits] [llvm/llvm-project] e47d5e: [AMDGPU] Hazard handling for gfx1250 wmma instruct...

[xur-llvm via All-commits]

  Branch: refs/heads/users/xur-llvm/icp_work
  Home:   https://github.com/llvm/llvm-project
  Commit: e47d5eb4541d5f377d9a57ef2157dbb3a41a85e6
      https://github.com/llvm/llvm-project/commit/e47d5eb4541d5f377d9a57ef2157dbb3a41a85e6
  Author: Changpeng Fang <changpeng.fang at amd.com>
  Date:   2025-07-21 (Mon, 21 Jul 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
    M llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h
    A llvm/test/CodeGen/AMDGPU/wmma-coececution-valu-hazards.mir
    A llvm/test/CodeGen/AMDGPU/wmma-hazards-gfx1250-w32.mir

  Log Message:
  -----------
  [AMDGPU] Hazard handling for gfx1250 wmma instructions (#149865)

If both instructions are xdl WMMA, hazard exists when the first WMMA
writes a register (D0) and the second WMMA reads it (A1/B1/Index1).

If the first instruction is a xdl WMMA, and the second one is a VALU,
three kinds of hazards exist:
  WMMA writes (D0), VALU reads (Use1);
  WMMA writes (D0), VALU writes (D1);
  WMMA reads (A0/B0.Index0), VALU writes (D1).

The actual number of hazard slots depends on the categories of the first
xdl WMMA as well as whether the second instruction is a xdl WMMA or
VALU. If there is not enough unrelated VALUs in between the two
instructions, appropriate number (to cover the missing) of V_NOPs will
be inserted to satisfy the hazard handling requirements.


  Commit: 2d31fc85a847759a5ea1fb39718a3851d57913c3
      https://github.com/llvm/llvm-project/commit/2d31fc85a847759a5ea1fb39718a3851d57913c3
  Author: Naveen Seth Hanig <naveen.hanig at outlook.com>
  Date:   2025-07-21 (Mon, 21 Jul 2025)

  Changed paths:
    M clang/include/clang/Basic/DiagnosticDriverKinds.td
    M clang/include/clang/Basic/DiagnosticGroups.td
    M clang/include/clang/Driver/Driver.h
    M clang/include/clang/Driver/Options.td
    M clang/include/clang/Lex/DependencyDirectivesScanner.h
    M clang/lib/Driver/Driver.cpp
    M clang/lib/Lex/DependencyDirectivesScanner.cpp
    A clang/test/Driver/modules-driver-cxx20-module-usage-scanner.cpp

  Log Message:
  -----------
  Reland [clang][modules-driver] Add scanner to detect C++20 module presence (#147630)

This patch is part of a series to natively support C++20 module usage
from the Clang driver (without requiring an external build system). This
introduces a new scanner that detects C++20 module usage in source files
without using the preprocessor or lexer.

For now, it is enabled only with the `-fmodules-driver` flag and serves
solely diagnostic purposes. In the future, the scanner will be enabled
for any (modules-driver compatible) compilation with two or more inputs,
and will help the driver determine whether to implicitly enable the
modules driver.

Since the scanner adds very little overhead, we are also exploring
enabling it for compilations with only a single input. This approach
could allow us to detect `import std` usage in a single-file
compilation, which would then activate the modules driver. For
performance measurements on this, see
https://github.com/naveen-seth/llvm-dev-cxx-modules-check-benchmark.

RFC:

https://discourse.llvm.org/t/rfc-modules-support-simple-c-20-modules-use-from-the-clang-driver-without-a-build-system

This patch relands commit ded1426. The CI failure is resolved by
removing the compatibility warning for using the `-fmodules-driver` flag
with pre-C++20 standards, which also better aligns its behavior with
other features/flags supported only in newer standards.


  Commit: d8ef5127f298f2f050cc1cd9b2f47e2f56f7fcb7
      https://github.com/llvm/llvm-project/commit/d8ef5127f298f2f050cc1cd9b2f47e2f56f7fcb7
  Author: xur-llvm <59886942+xur-llvm at users.noreply.github.com>
  Date:   2025-07-21 (Mon, 21 Jul 2025)

  Changed paths:
    M clang/include/clang/Basic/DiagnosticDriverKinds.td
    M clang/include/clang/Basic/DiagnosticGroups.td
    M clang/include/clang/Driver/Driver.h
    M clang/include/clang/Driver/Options.td
    M clang/include/clang/Lex/DependencyDirectivesScanner.h
    M clang/lib/Driver/Driver.cpp
    M clang/lib/Lex/DependencyDirectivesScanner.cpp
    A clang/test/Driver/modules-driver-cxx20-module-usage-scanner.cpp
    M llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
    M llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h
    A llvm/test/CodeGen/AMDGPU/wmma-coececution-valu-hazards.mir
    A llvm/test/CodeGen/AMDGPU/wmma-hazards-gfx1250-w32.mir

  Log Message:
  -----------
  Merge branch 'main' into users/xur-llvm/icp_work


Compare: https://github.com/llvm/llvm-project/compare/44f510b3f7d2...d8ef5127f298

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