metal : more precise Q*K in FA vec kernel #16558

Sign in to view logs

Triggered via pull request November 10, 2024 14:55

ggerganov

opened #10247

gg/metal-fa-f32-dot

Status Success

Total duration 54m 34s

Artifacts –

build.yml

on: pull_request

Matrix: windows-latest-cmake-cuda

Matrix: windows-latest-cmake-hip-release

Matrix: windows-latest-cmake

macOS-latest-cmake-arm64

macOS-latest-cmake-x64

ubuntu-focal-make

ubuntu-latest-cmake

macOS-latest-make

macOS-latest-cmake

ubuntu-focal-make-curl

ubuntu-latest-cmake-rpc

ubuntu-22-cmake-vulkan

ubuntu-22-cmake-hip

ubuntu-22-cmake-sycl

ubuntu-22-cmake-sycl-fp16

macOS-latest-cmake-ios

macOS-latest-cmake-tvos

windows-latest-cmake-sycl

windows-latest-cmake-hip

ios-xcode-build

Matrix: macOS-latest-swift

Matrix: ubuntu-latest-cmake-sanitizer

Matrix: windows-msys2

Annotations

1 error and 1 warning

windows-latest-cmake (avx512-x64, -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_...

Process completed with exit code 1.

android-build

The following actions use a deprecated Node.js version and will be forced to run on node20: actions/setup-java@v3. For more info: https://github.blog/changelog/2024-03-07-github-actions-all-actions-will-run-on-node20-instead-of-node16-by-default/