Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ci : move more jobs to ccache-buckets (#28375)

Browse Source 
* move more jobs to ccache-buckets

* add venv deps

* also jq
This commit is contained in:
Sigbjørn Skjæret
2026-09-04 15:50:33 +02:00
committed by GitHub
parent 1863ac0333
commit 8b4b3558f1
9 changed files with 230 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/build-apple.yml
+44 -24
View File
@@ -50,8 +50,16 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: apple-arm64
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: apple-arm64
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
@@ -67,6 +75,18 @@ jobs:
-DCMAKE_OSX_DEPLOYMENT_TARGET=13.3
time cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: apple-arm64
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Check for leaks
run: |
cmd=(./build/bin/test-thread-safety -hf ggml-org/gemma-3-270m-qat-GGUF -ngl 99 -p "$(printf 'hello %.0s' {1..128})" -n 16 -c 512 -ub 32 -np 2 -t 2 -lv 1)
@@ -80,16 +100,6 @@ jobs:
cd build
ctest -L main -E "test-llama-archs" --verbose --timeout 900
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: apple-arm64
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
macos-latest-x64:
runs-on: macos-15-intel
@@ -102,8 +112,16 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: apple-x64
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: apple-x64
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
@@ -120,22 +138,24 @@ jobs:
-DCMAKE_OSX_DEPLOYMENT_TARGET=13.3
time cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: apple-x64
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Test
id: cmake_test
run: |
cd build
ctest -L main --verbose --timeout 900
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: apple-x64
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
macos-latest-ios-xcode:
runs-on: macos-latest
.github/workflows/build-cpu.yml
+22 -14
View File
@@ -65,8 +65,7 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: cpu-${{ matrix.os }}
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: Build Dependencies
id: build_depends
@@ -91,6 +90,15 @@ jobs:
python3 -m pip install --upgrade pip setuptools
pip3 install ./gguf-py
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: cpu-${{ matrix.os }}
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
run: |
@@ -100,6 +108,18 @@ jobs:
-DGGML_RPC=ON
time cmake --build build --config Release -j $(nproc)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: cpu-${{ matrix.os }}
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Test
id: cmake_test
run: |
@@ -117,18 +137,6 @@ jobs:
./bin/llama-convert-llama2c-to-ggml --copy-vocab-from-model ./tok512.bin --llama2c-model stories260K.bin --llama2c-output-model stories260K.gguf
./bin/llama-completion -m stories260K.gguf -p "One day, Lily met a Shoggoth" -n 500 -c 256
# note: real deletion only on push to master (same condition as the ccache save),
# dry-run otherwise (the token is read-only on PRs from forks)
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: cpu-${{ matrix.os }}
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
windows:
name: windows / ${{ matrix.build }}
runs-on: windows-2025
.github/workflows/build-cuda-ubuntu.yml
+3 -3
View File
@@ -61,7 +61,7 @@ jobs:
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: cuda-ubuntu-24.04-cuda
folder: llama.cpp
@@ -116,7 +116,7 @@ jobs:
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: cuda-ubuntu-22.04-hip
folder: llama.cpp
@@ -167,7 +167,7 @@ jobs:
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: cuda-ubuntu-22.04-musa
folder: llama.cpp
.github/workflows/build-sycl.yml
+18 -8
View File
@@ -78,8 +78,16 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: sycl-ubuntu-24-${{ matrix.build }}
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: sycl-ubuntu-24-${{ matrix.build }}
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
@@ -96,15 +104,17 @@ jobs:
-DGGML_SYCL_F16=${{ matrix.fp16 }}
time cmake --build build --config Release -j $(nproc)
- name: ccache-clear
uses: ./.github/actions/ccache-clear
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
GH_TOKEN: ${{ github.token }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: sycl-ubuntu-24-${{ matrix.build }}
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
windows-latest-sycl:
runs-on: windows-2022
.github/workflows/build-vulkan.yml
+40 -20
View File
@@ -57,8 +57,16 @@ jobs:
with:
key: vulkan-ubuntu-24.04-arm
variant: ccache
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: vulkan-ubuntu-24.04-arm
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Configure
id: cmake_configure
@@ -73,15 +81,17 @@ jobs:
run: |
time cmake --build build -j $(nproc)
- name: ccache-clear
uses: ./.github/actions/ccache-clear
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
GH_TOKEN: ${{ github.token }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: vulkan-ubuntu-24.04-arm
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
ubuntu-llvmpipe:
runs-on: ubuntu-24.04
@@ -115,8 +125,16 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: vulkan-ubuntu-24.04-llvmpipe
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: vulkan-ubuntu-24.04-llvmpipe
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
@@ -127,6 +145,18 @@ jobs:
-DGGML_VULKAN=ON
cmake --build build --config Release -j $(nproc)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: vulkan-ubuntu-24.04-llvmpipe
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Test
id: cmake_test
run: |
@@ -138,16 +168,6 @@ jobs:
# test-backend-ops is too slow on llvmpipe, skip it
ctest -L main -E test-backend-ops --verbose --timeout 900
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: vulkan-ubuntu-24.04-llvmpipe
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
windows:
runs-on: windows-2025
.github/workflows/build-wasm.yml
+18 -8
View File
@@ -57,8 +57,7 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: webgpu-ubuntu-24.04-arm-wasm
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: Install Emscripten
run: |
@@ -76,6 +75,15 @@ jobs:
"https://github.com/google/dawn/releases/download/${DAWN_TAG}/${EMDAWN_PKG}"
unzip emdawn.zip
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: webgpu-ubuntu-24.04-arm-wasm
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build WASM WebGPU
run: |
source emsdk/emsdk_env.sh
@@ -89,12 +97,14 @@ jobs:
time cmake --build build-wasm --config Release --target test-backend-ops -j $(nproc)
- name: ccache-clear
uses: ./.github/actions/ccache-clear
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
GH_TOKEN: ${{ github.token }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: webgpu-ubuntu-24.04-arm-wasm
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
.github/workflows/build-webgpu.yml
+44 -24
View File
@@ -72,8 +72,7 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: webgpu-macos-latest
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: Dawn Dependency
id: dawn-depends
@@ -88,6 +87,15 @@ jobs:
mkdir dawn
tar -xvf artifact.tar.gz -C dawn --strip-components=1
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: webgpu-macos-latest
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
run: |
@@ -95,22 +103,24 @@ jobs:
cmake -B build -G "Ninja" -DCMAKE_BUILD_TYPE=Release -DGGML_WEBGPU=ON -DGGML_METAL=OFF -DGGML_BLAS=OFF
time cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: webgpu-macos-latest
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Test
id: cmake_test
run: |
cd build
ctest -L main --verbose --timeout 900
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: webgpu-macos-latest
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
ubuntu:
runs-on: ubuntu-24.04
@@ -123,8 +133,7 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: webgpu-ubuntu-24.04
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: Dependencies
id: depends
@@ -148,6 +157,15 @@ jobs:
mkdir dawn
tar -xvf artifact.tar.gz -C dawn --strip-components=1
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: webgpu-ubuntu-24.04
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
run: |
@@ -156,6 +174,18 @@ jobs:
-DGGML_WEBGPU=ON
time cmake --build build --config Release -j $(nproc)
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: webgpu-ubuntu-24.04
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Test
id: cmake_test
run: |
@@ -163,13 +193,3 @@ jobs:
# This is using llvmpipe and runs slower than other backends
# test-backend-ops is too slow on llvmpipe, skip it
ctest -L main -E test-backend-ops --verbose --timeout 900
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: webgpu-ubuntu-24.04
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
.github/workflows/hip-quality-check.yml
+19 -9
View File
@@ -49,14 +49,22 @@ jobs:
id: depends
run: |
sudo apt-get update
sudo apt-get install -y build-essential git cmake rocblas-dev hipblas-dev libssl-dev python3
sudo apt-get install -y build-essential git cmake rocblas-dev hipblas-dev libssl-dev python3 python3-venv python3-pip jq
- name: ccache
uses: ggml-org/ccache-action@v1.2.24
with:
key: hip-quality-check-ubuntu-22.04
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: hip-quality-check-ubuntu-22.04
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build with Werror
id: cmake_build
@@ -85,12 +93,14 @@ jobs:
make -j $(nproc) 2>&1 | tee metrics.log | grep -v 'Rpass-analysis=kernel-resource-usage\|remark:\|^$'
python3 ../scripts/hip/gcn-cdna-vgpr-check.py metrics.log
- name: ccache-clear
uses: ./.github/actions/ccache-clear
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
GH_TOKEN: ${{ github.token }}
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: hip-quality-check-ubuntu-22.04
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
.github/workflows/server.yml
+22 -12
View File
@@ -83,8 +83,16 @@ jobs:
uses: ggml-org/ccache-action@v1.2.24
with:
key: server-ubuntu-24.04-arm
evict-old-files: 1d
save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
save: false
- name: ccache-buckets-restore
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
with:
key: server-ubuntu-24.04-arm
folder: llama.cpp
hf_bucket: ggml-org/cache
- name: Build
id: cmake_build
@@ -93,6 +101,18 @@ jobs:
-DGGML_SCHED_NO_REALLOC=ON
cmake --build build --config Release -j $(nproc) --target llama-server
- name: ccache-buckets-save
if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
uses: ./.github/actions/ccache-buckets
env:
HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
with:
key: server-ubuntu-24.04-arm
folder: llama.cpp
evict-old-files: 1d
hf_bucket: ggml-org/cache
save: true
- name: Python setup
id: setup_python
uses: actions/setup-python@v6
@@ -128,16 +148,6 @@ jobs:
export LLAMA_ARG_BACKEND_SAMPLING=1
SLOW_TESTS=1 ./tests.sh
- name: ccache-clear
uses: ./.github/actions/ccache-clear
env:
GH_TOKEN: ${{ github.token }}
with:
key: server-ubuntu-24.04-arm
older: 5m
min: 1
dry-run: ${{ github.event_name != 'push' || github.ref != 'refs/heads/master' }}
windows:
runs-on: windows-2025