[ZenCPU] Update device to "zen5"

Signed-off-by: Chinmay Kulkarni <Chinmay.Kulkarni@amd.com>

.buildkite/hardware_tests/cpu.yaml

@@ -22,6 +22,29 @@ steps:
       pytest -x -v -s tests/kernels/test_onednn.py
       pytest -x -v -s tests/kernels/test_awq_int4_to_int8.py"
 
+- label: AMD-CPU-Kernel Tests
+  depends_on: []
+  soft_fail: false
+  device: zen5
+  no_plugin: true
+  source_file_dependencies:
+  - setup.py
+  - vllm/docker/Dockerfile.cpu
+  - vllm/requirements/cpu.txt
+  - vllm/requirements/build/cpu.txt
+  - csrc/cpu/
+  - cmake/cpu_extension.cmake
+  - CMakeLists.txt
+  - vllm/model_executor/layers/utils.py
+  - vllm/platforms/cpu.py
+  - vllm/platforms/zen_cpu.py
+  - vllm/platforms/__init__.py
+  - tests/model_executor/test_cpu_unquantized_gemm_dispatch.py
+  commands:
+    - |
+      bash .buildkite/scripts/hardware_ci/run-amd-cpu-test.sh 20m "
+      pytest -x -v -s tests/model_executor/test_cpu_unquantized_gemm_dispatch.py"
+
 - label: CPU-Compatibility Tests
   depends_on: []
   device: intel_cpu
@@ -35,6 +58,26 @@ steps:
       bash .buildkite/scripts/hardware_ci/run-cpu-test.sh 20m "
       bash .buildkite/scripts/hardware_ci/run-cpu-compatibility-test.sh"
 
+- label: AMD-CPU-Compatibility Tests
+  depends_on: []
+  soft_fail: false
+  device: zen5
+  no_plugin: true
+  source_file_dependencies:
+  - setup.py
+  - vllm/docker/Dockerfile.cpu
+  - vllm/requirements/cpu.txt
+  - vllm/requirements/build/cpu.txt
+  - vllm/platforms/cpu.py
+  - vllm/platforms/zen_cpu.py
+  - vllm/platforms/interface.py
+  - vllm/platforms/__init__.py
+  - tests/test_zen_cpu_platform_detection.py
+  commands:
+    - |
+      bash .buildkite/scripts/hardware_ci/run-amd-cpu-test.sh 20m "
+      pytest -x -v -s tests/test_zen_cpu_platform_detection.py"
+
 - label: CPU-Language Generation and Pooling Model Tests
   depends_on: []
   device: intel_cpu
@@ -50,6 +93,32 @@ steps:
       pytest -x -v -s tests/models/language/generation -m cpu_model
       pytest -x -v -s tests/models/language/pooling -m cpu_model"
 
+- label: AMD-CPU-Language Generation and Pooling Model Tests
+  depends_on: []
+  soft_fail: false
+  device: zen5
+  no_plugin: true
+  source_file_dependencies:
+  - setup.py
+  - vllm/docker/Dockerfile.cpu
+  - vllm/requirements/cpu.txt
+  - vllm/requirements/build/cpu.txt
+  - csrc/cpu/
+  - vllm/model_executor/layers/utils.py
+  - vllm/platforms/zen_cpu.py
+  - vllm/platforms/__init__.py
+  - setup.py
+  - vllm/platforms/cpu.py
+  - vllm/platforms/interface.py
+  - vllm/v1/worker/cpu_model_runner.py
+  - tests/models/language/generation/
+  - tests/models/language/pooling/
+  commands:
+    - |
+      bash .buildkite/scripts/hardware_ci/run-amd-cpu-test.sh 30m "
+      pytest -x -v -s tests/models/language/generation -m cpu_model
+      pytest -x -v -s tests/models/language/pooling -m cpu_model"
+
 - label: CPU-Quantization Model Tests
   depends_on: []
   device: intel_cpu
@@ -98,6 +167,31 @@ steps:
       bash .buildkite/scripts/hardware_ci/run-cpu-test.sh 10m "
       bash .buildkite/scripts/hardware_ci/run-cpu-distributed-smoke-test.sh dp_tp"
 
+- label: AMD-CPU-Distributed Tests
+  depends_on: []
+  soft_fail: false
+  device: zen5
+  no_plugin: true
+  source_file_dependencies:
+  - setup.py
+  - vllm/docker/Dockerfile.cpu
+  - vllm/requirements/cpu.txt
+  - vllm/requirements/build/cpu.txt
+  - csrc/cpu/shm.cpp
+  - vllm/v1/worker/cpu_worker.py
+  - vllm/v1/worker/gpu_worker.py
+  - vllm/v1/worker/cpu_model_runner.py
+  - vllm/v1/worker/gpu_model_runner.py
+  - vllm/platforms/cpu.py
+  - vllm/platforms/zen_cpu.py
+  - vllm/platforms/cpu.py
+  - vllm/distributed/parallel_state.py
+  - vllm/distributed/device_communicators/cpu_communicator.py
+  commands:
+    - |
+      bash .buildkite/scripts/hardware_ci/run-amd-cpu-test.sh 10m "
+      bash .buildkite/scripts/hardware_ci/run-cpu-distributed-smoke-test.sh"
+
 - label: CPU-Multi-Modal Model Tests %N
   depends_on: []
   device: intel_cpu

.buildkite/image_build/image_build.yaml

@@ -27,6 +27,21 @@ steps:
         - exit_status: -10  # Agent was lost
           limit: 2
 
+  - label: ":docker: Build AMD CPU image"
+    soft_fail: true
+    key: image-build-amd-cpu
+    depends_on: []
+    commands:
+    - .buildkite/image_build/image_build_amd_cpu.sh $REGISTRY $REPO $BUILDKITE_COMMIT
+    env:
+      DOCKER_BUILDKIT: "1"
+    retry:
+      automatic:
+        - exit_status: -1  # Agent was lost
+          limit: 2
+        - exit_status: -10  # Agent was lost
+          limit: 2
+
   - label: ":docker: Build HPU image"
     soft_fail: true
     depends_on: []

.buildkite/image_build/image_build_amd_cpu.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+set -e
+
+if [[ $# -lt 3 ]]; then
+  echo "Usage: $0 <registry> <repo> <commit>"
+  exit 1
+fi
+
+REGISTRY=$1
+REPO=$2
+BUILDKITE_COMMIT=$3
+
+# authenticate with AWS ECR
+aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin "$REGISTRY"
+
+# skip build if image already exists
+if [[ -z $(docker manifest inspect "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-amd-cpu) ]]; then
+  echo "Image not found, proceeding with build..."
+else
+  echo "Image found"
+  exit 0
+fi
+
+# build
+docker build --file docker/Dockerfile.cpu \
+  --build-arg max_jobs=16 \
+  --build-arg buildkite_commit="$BUILDKITE_COMMIT" \
+  --build-arg VLLM_CPU_X86=true \
+  --tag "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-amd-cpu \
+  --target vllm-zen-test \
+  --progress plain .
+
+# push
+docker push "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-amd-cpu

.buildkite/scripts/hardware_ci/run-amd-cpu-test.sh

@@ -0,0 +1,20 @@
+#!/bin/bash
+
+# This script build the CPU docker image and run the offline inference inside the container.
+# It serves a sanity check for compilation and basic model usage.
+set -euox pipefail
+
+# allow to bind to different cores
+CORE_RANGE=${CORE_RANGE:-48-95}
+NUMA_NODE=${NUMA_NODE:-1}
+IMAGE_NAME="amd-cpu-test-$NUMA_NODE"
+TIMEOUT_VAL=$1
+TEST_COMMAND=$2
+
+# building the docker image
+echo "--- :docker: Building Docker image"
+docker build --progress plain --tag "$IMAGE_NAME" --target vllm-zen-test -f docker/Dockerfile.cpu .
+
+# Run the image, setting --shm-size=4g for tensor parallel.
+docker run --rm --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN -e VLLM_CPU_KVCACHE_SPACE=16 -e VLLM_CPU_CI_ENV=1 -e VLLM_CPU_SIM_MULTI_NUMA=1 --shm-size=4g "$IMAGE_NAME" \
+        timeout "$TIMEOUT_VAL" bash -c "set -euox pipefail; echo \"--- Print packages\"; pip list; echo \"--- Running tests\"; ${TEST_COMMAND}"

docker/Dockerfile.cpu

@@ -250,3 +250,22 @@ RUN --mount=type=cache,target=/root/.cache/uv \
     uv pip install "vllm[zen]"
 
 ENTRYPOINT ["vllm", "serve"]
+
+######################### ZEN CPU TEST IMAGE #########################
+FROM vllm-openai-zen AS vllm-zen-test
+
+COPY --from=vllm-test-deps /vllm-workspace/requirements/cpu-test.txt requirements/test.txt
+
+RUN --mount=type=cache,target=/root/.cache/uv \
+    uv pip install -r requirements/test.txt
+
+ADD ./tests/ ./tests/
+ADD ./examples/ ./examples/
+ADD ./benchmarks/ ./benchmarks/
+ADD ./vllm/collect_env.py .
+ADD ./.buildkite/ ./.buildkite/
+
+RUN --mount=type=cache,target=/root/.cache/uv \
+    uv pip install -e tests/vllm_test_utils
+
+ENTRYPOINT []

tests/models/language/generation/test_common.py

@@ -42,6 +42,11 @@ AITER_MODEL_LIST = [
                 pytest.mark.core_model,
                 pytest.mark.slow_test,
                 pytest.mark.cpu_model,
+                pytest.mark.skipif(
+                    current_platform.is_zen_cpu(),
+                    reason="bloom-560m ALiBi is currently not supported on\
+                        AMD Zen CPUs due to lack of support for float16 compute.",
+                ),
             ],
         ),
         pytest.param(

tests/models/language/pooling/test_reward.py

@@ -88,7 +88,15 @@ def load_reward_outputs(filename: "StrPath") -> list[list[float]]:
     [
         pytest.param(
             "Qwen/Qwen2.5-Math-PRM-7B",
-            marks=[pytest.mark.core_model, pytest.mark.cpu_model],
+            marks=[
+                pytest.mark.core_model,
+                pytest.mark.cpu_model,
+                pytest.mark.skipif(
+                    current_platform.is_zen_cpu(),
+                    reason="Qwen2.5-Math-PRM-7B is currently not supported on\
+                        AMD Zen CPUs due to lack of support for float16 compute.",
+                )
+            ],
         ),
     ],
 )
@@ -131,7 +139,15 @@ def test_prm_models(
     [
         pytest.param(
             "Qwen/Qwen2.5-Math-PRM-7B",
-            marks=[pytest.mark.core_model, pytest.mark.cpu_model],
+            marks=[
+                pytest.mark.core_model,
+                pytest.mark.cpu_model,
+                pytest.mark.skipif(
+                    current_platform.is_zen_cpu(),
+                    reason="Qwen2.5-Math-PRM-7B is currently not supported on\
+                        AMD Zen CPUs due to lack of support for float16 compute.",
+                )
+            ],
         ),
     ],
 )