From d903abfccc0f0fc8e73364cf5418a26118bda99e Mon Sep 17 00:00:00 2001
From: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
Date: Fri, 17 Nov 2023 10:44:37 +0100
Subject: [PATCH] Fix AMD CI not showing GPU (#27555)

fix

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
---
 .github/workflows/self-push-amd.yml | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/.github/workflows/self-push-amd.yml b/.github/workflows/self-push-amd.yml
index c72f224a30..19857981b1 100644
--- a/.github/workflows/self-push-amd.yml
+++ b/.github/workflows/self-push-amd.yml
@@ -38,14 +38,16 @@ jobs:
     runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
     container:
       image: huggingface/transformers-pytorch-amd-gpu-push-ci  # <--- We test only for PyTorch for now
-      options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
       - name: ROCM-SMI
         run: |
-          rocminfo  | grep "Agent" -A 14
-      - name: Show HIP environment
+          rocm-smi
+      - name: ROCM-INFO
+        run: |
+          rocminfo  | grep "Agent" -A 14
+      - name: Show ROCR environment
         run: |
-          echo "HIP: $HIP_VISIBLE_DEVICES"
           echo "ROCR: $ROCR_VISIBLE_DEVICES"
 
   setup_gpu:
@@ -57,7 +59,7 @@ jobs:
     runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
     container:
       image: huggingface/transformers-pytorch-amd-gpu-push-ci  # <--- We test only for PyTorch for now
-      options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     outputs:
       matrix: ${{ steps.set-matrix.outputs.matrix }}
       test_map: ${{ steps.set-matrix.outputs.test_map }}
@@ -155,7 +157,7 @@ jobs:
     runs-on: [self-hosted, docker-gpu, amd-gpu, '${{ matrix.machine_type }}', '${{ inputs.gpu_flavor }}']
     container:
       image: huggingface/transformers-pytorch-amd-gpu-push-ci  # <--- We test only for PyTorch for now
-      options: --device /dev/kfd --device /dev/dri --env HIP_VISIBLE_DEVICES --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --device /dev/kfd --device /dev/dri --env ROCR_VISIBLE_DEVICES --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
       # Necessary to get the correct branch name and commit SHA for `workflow_run` event
       # We also take into account the `push` event (we might want to test some changes in a branch)
@@ -207,10 +209,12 @@ jobs:
 
       - name: ROCM-SMI
         run: |
-          rocminfo  | grep "Agent" -A 14
-      - name: Show HIP environment
+          rocm-smi
+      - name: ROCM-INFO
+        run: |
+          rocminfo  | grep "Agent" -A 14
+      - name: Show ROCR environment
         run: |
-          echo "HIP: $HIP_VISIBLE_DEVICES"
           echo "ROCR: $ROCR_VISIBLE_DEVICES"
 
       - name: Environment