Trim macos jobs and do as much as possible in aarch64 instances

Author: mergennachin

.github/workflows/pull.yml

@@ -49,11 +49,20 @@ jobs:
         model: [mv3, vit]
         backend: [portable, xnnpack-quantization-delegation]
         build-tool: [cmake, buck2]
-        runner: [linux.2xlarge]
+        runner: [linux.2xlarge, linux.arm64.2xlarge]
+        docker-image: [executorch-ubuntu-22.04-clang12, executorch-ubuntu-22.04-gcc11-aarch64]
+        # Excluding specific runner + docker image combinations that don't make sense:
+        #   - Excluding the ARM64 gcc image on the x86 runner (linux.2xlarge)
+        #   - Excluding the x86 clang image on the ARM64 runner (linux.arm64.2xlarge)
+        exclude:
+          - runner: linux.2xlarge
+            docker-image: executorch-ubuntu-22.04-gcc11-aarch64
+          - runner: linux.arm64.2xlarge
+            docker-image: executorch-ubuntu-22.04-clang12
       fail-fast: false
     with:
       runner: ${{ matrix.runner }}
-      docker-image: executorch-ubuntu-22.04-clang12
+      docker-image: ${{ matrix.docker-image }}
       submodules: 'true'
       ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
       timeout: ${{ matrix.timeout }}

.github/workflows/trunk.yml

@@ -23,25 +23,21 @@ jobs:
     uses: pytorch/test-infra/.github/workflows/macos_job.yml@main
     strategy:
       matrix:
-        model: [add, add_mul, emformer_join, emformer_transcribe, ic3, ic4, linear, llama2, mobilebert, mv2, mv3, resnet18, resnet50, vit, w2l]
-        backend: [portable, xnnpack-quantization-delegation]
+        model: [emformer_join, ic4, llama2, mobilebert, mv3, resnet50, vit, w2l]
+        backend: [xnnpack-quantization-delegation]
         include:
           - model: efficient_sam
             backend: portable
           - model: llama
             backend: portable
           - model: llama3_2_vision_encoder
             backend: portable
-          - model: lstm
-            backend: portable
-          - model: mul
+          - model: mv3
             backend: portable
           - model: phi-4-mini
             backend: portable
           - model: qwen2_5
             backend: portable
-          - model: softmax
-            backend: portable
       fail-fast: false
     with:
       runner: macos-m1-stable
@@ -71,10 +67,16 @@ jobs:
       matrix:
         model: [linear, add, add_mul, ic3, ic4, mv2, mv3, resnet18, resnet50, vit, w2l, mobilebert, emformer_join, emformer_transcribe]
         backend: [portable, xnnpack-quantization-delegation]
-        runner: [linux.arm64.2xlarge]
+        include:
+          - model: lstm
+            backend: portable
+          - model: mul
+            backend: portable
+          - model: softmax
+            backend: portable
       fail-fast: false
     with:
-      runner: ${{ matrix.runner }}
+      runner: linux.arm64.2xlarge
       docker-image: executorch-ubuntu-22.04-gcc11-aarch64
       submodules: 'true'
       ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}