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

[RELEASE] [AMD] Additional AMD cherry-picks #4175

Merged
merged 7 commits into from
Jun 20, 2024
Merged

[RELEASE] [AMD] Additional AMD cherry-picks #4175

merged 7 commits into from
Jun 20, 2024

Commits on Jun 19, 2024

  1. [BACKEND][AMD] Disable linear layout due to perf regression (triton-l… 

    …ang#4126)

We have identified a 20% perf regression in our downstream flash
attention perf kernel after switching to linear layout. Initial analysis
shows register pressure is increased to cause spills. Further analysis
is still ongoing.

So this commit introduces a minimal way to selectively disable linear
layout only on AMD backend to avoid affecting NVIDIA backend while
continuing bring it up on AMD side.

(cherry picked from commit e8bc45d)
    @antiagainst @jataylo
    antiagainst authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    5865d77 View commit details
    Browse the repository at this point in the history

  2. [AMD] Add basic verification to MFMA encoding (triton-lang#4117) 

    This guards us against unsupported cases without asserts. Along the way
slightly improved the MFMA/WMMA doc a bit.

(cherry picked from commit 9a0a7c2)
    @antiagainst @jataylo
    antiagainst authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    41948fb View commit details
    Browse the repository at this point in the history

  3. [AMD][WMMA] Support dot3d (triton-lang#3674) 

    This PR enables support of 3d dot for RDNA GPUs.

(cherry picked from commit 100e2aa)
    @binarman @jataylo
    binarman authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    908a08d View commit details
    Browse the repository at this point in the history

  4. [AMD][gfx11] Fix BF16 wmma instr generation (triton-lang#4135) 

    - Pack bf16 elements to int16 vectors;
- Add a lit test;
- BF16 testcases from test_core.py::test_dot are passed for now;

The core Triton is a small number of people, and we receive many PRs
(thank
you!).  To help us review your code more quickly, **if you are a new
contributor (less than 3 PRs merged) we ask that you complete the
following
tasks and include the filled-out checklist in your PR description.**

Complete the following tasks before sending your PR, and replace `[ ]`
with
`[x]` to indicate you have done them.

- [x] I am not making a trivial change, such as fixing a typo in a
comment.

- [x] I have written a PR description following these
  [rules](https://cbea.ms/git-commit/#why-not-how).

- [x] I have run `pre-commit run --from-ref origin/main --to-ref HEAD`.

- Select one of the following.
  - [x] I have added tests.
    - `/test` for `lit` tests

- Select one of the following.
- [x] The `lit` tests I have added follow these [best
practices](https://mlir.llvm.org/getting_started/TestingGuide/#filecheck-best-practices),
including the "tests should be minimal" section. (Usually running Python
code
    and using the instructions it generates is not minimal.)

Signed-off-by: Ilya Veselov <iveselov.nn@gmail.com>
(cherry picked from commit 4a1ea8e)
    @joviliast @jataylo
    joviliast authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    821e258 View commit details
    Browse the repository at this point in the history

  5. [PROTON] Refactor GPU profilers (triton-lang#4056) 

    1. Extract duplicated code into GPUProfiler.h
2. Track finished correlation ids for both cupti and amd profilers

(cherry picked from commit 328b86d)
    @Jokeren @jataylo
    Jokeren authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    06dc808 View commit details
    Browse the repository at this point in the history

  6. [PROTON] Roctracer: convert agent id to gpu id for gpu ops (triton-la… 

    …ng#4090)

Roctracer reports (global) agent ids for the location of async ops, e.g.
kernels and copies.
The profiler would be better suited with gpu indexes (zero based).

Created a mapping function to apply to values stored in
KernelMetric::DeviceId.

Caveat: if devices are hidden using HIP_VISIBLE_DEVICES then the hip
device id, e.g. via hipGetDevice()/hipSetDevice(), will not match the
reported unfiltered id. Additional support in hip will be needed to map
through the filtering correctly.

---------

Co-authored-by: Keren Zhou <robinho364@gmail.com>
(cherry picked from commit 60613fb)
    @mwootton @jataylo
    mwootton authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    6c94213 View commit details
    Browse the repository at this point in the history

  7. [PROTON][AMD] Add Proton HIP GPU Utilization Metrics (triton-lang#4119) 

    This PR adds Proton HIP GPU utilization metrics and an associated test.

(cherry picked from commit c1776fa)
    @CRobeck @jataylo
    CRobeck authored and jataylo committed Jun 19, 2024
    Configuration menu
    Copy the full SHA
    f65efe4 View commit details
    Browse the repository at this point in the history