Fix vLLM benchmark setup on HPU runners (#100)

Signed-off-by: Huy Do <[email protected]>

Author: huydhn

.github/workflows/vllm-benchmark.yml

@@ -250,7 +250,7 @@ jobs:
 
           pushd vllm-benchmarks/vllm
           git checkout "${HEAD_SHA}"
-          rm .buildkite/performance-benchmarks/tests/*.json
+          rm .buildkite/performance-benchmarks/tests/*.json || true
           popd
 
           # Set the list of benchmarks we want to cover in this runner
@@ -310,7 +310,7 @@ jobs:
 
       - name: Authenticate with AWS
         # AWS CUDA runners already have access to the bucket via its runner IAM role
-        if: env.DEVICE_NAME == 'rocm' || contains(env.DEVICE_TYPE, 'B200')
+        if: env.DEVICE_NAME == 'rocm' || env.DEVICE_NAME == 'hpu' || contains(env.DEVICE_TYPE, 'B200') || contains(matrix.runner, 'gnr')
         uses: aws-actions/configure-aws-credentials@ececac1a45f3b08a01d2dd070d28d111c5fe6722 # v4.1.0
         with:
           role-to-assume: arn:aws:iam::308535385114:role/gha_workflow_upload-benchmark-results