A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 5, 23:30 UTC
[cpu][ci] remove soft-fail for Arm CI and add quant model tests (#37691)
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
Fadi Arafeh committed
Mar 26, 2026 at 07:03 UTC
71161e8b63f9534d6ac5e098a4874621164d1f1e
2 files changed
+8
-2
.buildkite/hardware_tests/cpu.yaml
+1
@@ -101,6 +101,7 @@ steps:
101
102
- label: "Arm CPU Test"
103
depends_on: []
104
+ soft_fail: false
105
device: arm_cpu
106
no_plugin: true
107
commands:
.buildkite/scripts/hardware_ci/run-cpu-test-arm.sh
+7
-2
@@ -5,8 +5,8 @@
5
set -ex
6
7
# allow to bind to different cores
8
-CORE_RANGE=${CORE_RANGE:-0-16}
9
-OMP_CORE_RANGE=${OMP_CORE_RANGE:-0-16}
8
+CORE_RANGE=${CORE_RANGE:-0-31}
9
+OMP_CORE_RANGE=${OMP_CORE_RANGE:-0-31}
10
11
export CMAKE_BUILD_PARALLEL_LEVEL=16
12
@@ -41,6 +41,11 @@ function cpu_tests() {
41
set -e
42
pytest -x -v -s tests/models/multimodal/generation/test_whisper.py -m cpu_model"
43
44
+ # Run quantized model tests
45
+ docker exec cpu-test bash -c "
46
+ set -e
47
+ pytest -x -v -s tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_logprobs"
48
+
49
# Run kernel tests
50
docker exec cpu-test bash -c "
51
set -e