3.3. AMD GPU Operator 테스트

프로세스

1. ROCmInfo를 테스트하는 YAML 파일을 생성합니다.

   $ cat << EOF > rocminfo.yaml
   
   apiVersion: v1
   kind: Pod
   metadata:
    name: rocminfo
   spec:
    containers:
    - image: docker.io/rocm/pytorch:latest
      name: rocminfo
      command: ["/bin/sh","-c"]
      args: ["rocminfo"]
      resources:
       limits:
         amd.com/gpu: 1
       requests:
         amd.com/gpu: 1
    restartPolicy: Never
   EOF

2. rocminfo Pod를 생성합니다.

   $ oc create -f rocminfo.yaml

   출력 예

   apiVersion: v1
   pod/rocminfo created

3. MI210 GPU 1개로 rocmnfo 로그를 확인합니다.

   $ oc logs rocminfo | grep -A5 "Agent"

   출력 예

   HSA Agents
   ==========
   *******
   Agent 1
   *******
     Name:                    Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz
     Uuid:                    CPU-XX
     Marketing Name:          Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz
     Vendor Name:             CPU
   --
   Agent 2
   *******
     Name:                    Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz
     Uuid:                    CPU-XX
     Marketing Name:          Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz
     Vendor Name:             CPU
   --
   Agent 3
   *******
     Name:                    gfx90a
     Uuid:                    GPU-024b776f768a638b
     Marketing Name:          AMD Instinct MI210
     Vendor Name:             AMD

4. Pod를 삭제합니다.

   $ oc delete -f rocminfo.yaml

   출력 예

   pod "rocminfo" deleted