Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[AMD] Reland instruction scheduling hint changes #4940

Merged
merged 12 commits into from
Oct 31, 2024
Merged

[AMD] Reland instruction scheduling hint changes #4940

merged 12 commits into from
Oct 31, 2024

Commits on Oct 29, 2024

  1. Revert "[AMD] revert optimizations (triton-lang#4919)" 

    This reverts commit 93de426.
    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    4b50c48 View commit details
    Browse the repository at this point in the history

  2. [AMD] use rocdl instr.sched.barriers from upstream MLIR/ROCDL

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    d53c499 View commit details
    Browse the repository at this point in the history

  3. [AMD] fixed a bug resulted in reverting PR#4919 

    Replaced temlate-based impl. of `rewindUnaryOps` in
`SchedInstructions.cpp` using regular for-loops.
The new impl. is more robust and can handle
other unary ops automatically.
    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    ce02968 View commit details
    Browse the repository at this point in the history

  4. [AMD] Moved annotateDotUsageOnLoadStore to stream pipeliner

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    2aecafa View commit details
    Browse the repository at this point in the history

  5. [AMD] Fixed bug in setNumGeneratedGlobalLoads 

    * add a test for the presence of OpIdx attribute
    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    088fbd9 View commit details
    Browse the repository at this point in the history

  6. [AMD] added additional check into createCKV3Schedule 

    The extra check tests whether the data are loaded from HBM
using `buffer_load` instructions. The CKV3 scheduling is
skipped if the check fails.
    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    c72bfb9 View commit details
    Browse the repository at this point in the history

  7. [AMD] Udated tests for SchedInstructions passes

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    68e7fac View commit details
    Browse the repository at this point in the history

  8. [AMD] Addressed comments of PR#4940

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    dd7d2c6 View commit details
    Browse the repository at this point in the history

  9. [AMD] Fixed propagation of OpIdx attribute in LoadToBufferLoad pass

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    a2f8874 View commit details
    Browse the repository at this point in the history

  10. [AMD] Fixed instruction.sched lit tests

    @ravil-mobile
    ravil-mobile committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    ae8c3c8 View commit details
    Browse the repository at this point in the history

  11. Drop unnecessary mlir::prefix and early return if none choice

    @antiagainst
    antiagainst committed Oct 29, 2024
    Configuration menu
    Copy the full SHA
    2504666 View commit details
    Browse the repository at this point in the history

Commits on Oct 31, 2024

  1. Merge branch 'main' into ravil/bug-fix

    @antiagainst
    antiagainst authored Oct 31, 2024
    Configuration menu
    Copy the full SHA
    63a552f View commit details
    Browse the repository at this point in the history