Skip to content

Commit 20c1857

Browse files
[SYCL][DevOps] Switch GEN9 Linux pre-commit to fused L0/OCL task (#9500)
SYCL End-to-End test spend most of the time on compilation. Running them against two targets with a single compilation can reduced the load generated in half.
1 parent f363bb2 commit 20c1857

File tree

2 files changed

+13
-1
lines changed

2 files changed

+13
-1
lines changed

.github/workflows/sycl_precommit.yml

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -48,7 +48,7 @@ jobs:
4848
uses: ./.github/workflows/sycl_gen_test_matrix.yml
4949
with:
5050
ref: ${{ github.event.pull_request.head.sha }}
51-
lts_config: "hip_amdgpu;ocl_x64;ocl_gen9;l0_gen9;esimd_emu;cuda_aws;win_l0_gen12"
51+
lts_config: "hip_amdgpu;ocl_x64;lin_gen9;esimd_emu;cuda_aws;win_l0_gen12"
5252

5353
linux_default:
5454
name: Linux

devops/test_configs.json

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,17 @@
11
{
22
"lts": [
3+
{
4+
"config": "lin_gen9",
5+
"name": "SYCL E2E on Intel GEN9 GPU",
6+
"runs-on": [
7+
"Linux",
8+
"gen9"
9+
],
10+
"image": "${{ inputs.intel_drivers_image }}",
11+
"container_options": "-u 1001 --device=/dev/dri --privileged --cap-add SYS_ADMIN",
12+
"targets": "ext_oneapi_level_zero:gpu;opencl:gpu",
13+
"cmake_args": ""
14+
},
315
{
416
"config": "l0_gen9",
517
"name": "L0 GEN9 LLVM Test Suite",

0 commit comments

Comments
 (0)