cpp_wrapper: enable in aarch64 and x86 nightly dashboard performance runs (pytorch#145791)

Adds `cpp_wrapper` mode to the nightly inductor benchmark runs, as well as optionally for manually triggered runs. This is justified by `aot_inductor` already being in those runs.

Additionally, re-enables `aot_inductor` in the nightly aarch64 runs. It was disabled 5 months ago to deal with a performance instability, which has likely gone away at this point.
Pull Request resolved: pytorch#145791
Approved by: https://github.com/desertfire

Author: benjaminglass1

.github/workflows/inductor-perf-test-nightly-aarch64.yml

@@ -2,8 +2,6 @@ name: inductor-perf-nightly-aarch64
 
 on:
   schedule:
-    # - cron: 0 7 * * 1-6
-    # - cron: 0 7 * * 0
     # Does not perform max_autotune on CPU, so skip the weekly run setup
     - cron: 0 7 * * *
   # NB: GitHub has an upper limit of 10 inputs here
@@ -30,6 +28,11 @@ on:
         required: false
         type: boolean
         default: false
+      cppwrapper:
+        description: Run inductor_cpp_wrapper?
+        required: false
+        type: boolean
+        default: false
       aotinductor:
         description: Run aot_inductor for inference?
         required: false
@@ -117,9 +120,7 @@ jobs:
     if: github.event.schedule == '0 7 * * *'
     with:
       build-environment: linux-jammy-aarch64-py3.10
-      # Turn off dynamic-shapes and aotinductor tests for now, to have faster iteration for debugging perf instability.
-      # Will change this back
-      dashboard-tag: training-false-inference-true-default-true-dynamic-false-aotinductor-false
+      dashboard-tag: training-false-inference-true-default-true-dynamic-true-cppwrapper-true-aotinductor-true
       docker-image: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.test-matrix }}
       timeout-minutes: 720
@@ -135,7 +136,7 @@ jobs:
     if: github.event_name == 'workflow_dispatch'
     with:
       build-environment: linux-jammy-aarch64-py3.10
-      dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-aotinductor-${{ inputs.aotinductor }}
+      dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-cppwrapper-${{ inputs.cppwrapper }}-aotinductor-${{ inputs.aotinductor }}
       docker-image: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.test-matrix }}
       timeout-minutes: 720

.github/workflows/inductor-perf-test-nightly-x86.yml

@@ -30,6 +30,11 @@ on:
         required: false
         type: boolean
         default: false
+      cppwrapper:
+        description: Run inductor_cpp_wrapper?
+        required: false
+        type: boolean
+        default: false
       aotinductor:
         description: Run aot_inductor for inference?
         required: false
@@ -92,7 +97,7 @@ jobs:
     if: github.event.schedule == '0 7 * * *'
     with:
       build-environment: linux-jammy-py3.9-gcc11-build
-      dashboard-tag: training-false-inference-true-default-true-dynamic-true-aotinductor-true
+      dashboard-tag: training-false-inference-true-default-true-dynamic-true-cppwrapper-true-aotinductor-true
       docker-image: ${{ needs.linux-jammy-cpu-py3_9-gcc11-inductor-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-jammy-cpu-py3_9-gcc11-inductor-build.outputs.test-matrix }}
       timeout-minutes: 720
@@ -108,7 +113,7 @@ jobs:
     if: github.event_name == 'workflow_dispatch'
     with:
       build-environment: linux-jammy-py3.9-gcc11-build
-      dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-aotinductor-${{ inputs.aotinductor }}
+      dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-cppwrapper-${{ inputs.cppwrapper }}-aotinductor-${{ inputs.aotinductor }}
       docker-image: ${{ needs.linux-jammy-cpu-py3_9-gcc11-inductor-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-jammy-cpu-py3_9-gcc11-inductor-build.outputs.test-matrix }}
       timeout-minutes: 720