[SYCL] Add Q8_0 reorder optimization for Intel GPUs (~3x token generation speedup) #38577
build.yml
on: pull_request
build-cmake-pkg
/
linux
8m 33s
macOS-latest-arm64
5m 43s
macOS-latest-x64
5m 18s
macOS-latest-arm64-webgpu
5m 29s
ubuntu-latest-rpc
5m 51s
ubuntu-24-webgpu
15m 38s
ubuntu-24-webgpu-wasm
3m 4s
ubuntu-22-hip
2m 19s
ubuntu-22-musa
20m 39s
ubuntu-22-sycl
18m 25s
ubuntu-22-sycl-fp16
10m 46s
ubuntu-latest-cuda
1m 42s
windows-latest-sycl
18m 30s
windows-latest-hip
25m 47s
ubuntu-cpu-riscv64-native
1h 2m
ggml-ci-x64-cpu-low-perf
4m 33s
ggml-ci-arm64-cpu-low-perf
4m 5s
ggml-ci-x64-cpu-high-perf
28m 51s
ggml-ci-arm64-cpu-high-perf
15m 52s
ggml-ci-arm64-cpu-high-perf-sve
15m 22s
ggml-ci-arm64-cpu-kleidiai
21m 24s
ggml-ci-arm64-cpu-kleidiai-graviton4
8m 47s
Matrix: ubuntu-24-openvino
Matrix: ubuntu-24-vulkan
Matrix: ubuntu-cpu
Matrix: windows-2022-cuda
Matrix: windows-latest