Add testing on A10G GPU to periodic workflow (pytorch#85524)

This enables testing on lots of modern CUDA features on sm_86 capable GPU

While migrating to that platform, discovered that `functorch` tests for `nn.functional.conv.transpose3d` produce garbage on sm_80+ as well as 2 `nvfuser` tests unexpectedly pass and one unexpectedly fails.

TODO:
 - Investigate unexpected success for `test_vmapvjp_linalg_householder_product_cuda_float32` and add `functorch` shard

Pull Request resolved: pytorch#85524
Approved by: https://github.com/ngimel

.github/workflows/periodic.yml

@@ -148,6 +148,30 @@ jobs:
       docker-image: ${{ needs.linux-bionic-cuda11_7-py3_7-gcc7-debug-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-bionic-cuda11_7-py3_7-gcc7-debug-build.outputs.test-matrix }}
 
+  linux-bionic-cuda11_6-py3_10-gcc7-sm86-build:
+    name: cuda11.6-py3.10-gcc7-sm86
+    uses: ./.github/workflows/_linux-build.yml
+    with:
+      build-environment: linux-bionic-cuda11.6-py3.10-gcc7-sm86
+      docker-image-name: pytorch-linux-bionic-cuda11.6-cudnn8-py3-gcc7
+      cuda-arch-list: 8.6
+      test-matrix: |
+        { include: [
+          { config: "default", shard: 1, num_shards: 4, runner: "linux.g5.4xlarge.nvidia.gpu" },
+          { config: "default", shard: 2, num_shards: 4, runner: "linux.g5.4xlarge.nvidia.gpu" },
+          { config: "default", shard: 3, num_shards: 4, runner: "linux.g5.4xlarge.nvidia.gpu" },
+          { config: "default", shard: 4, num_shards: 4, runner: "linux.g5.4xlarge.nvidia.gpu" },
+        ]}
+
+  linux-bionic-cuda11_6-py3_10-gcc7-sm86-test:
+    name: cuda11.6-py3.10-gcc7-sm86
+    uses: ./.github/workflows/_linux-test.yml
+    needs: linux-bionic-cuda11_6-py3_10-gcc7-sm86-build
+    with:
+      build-environment: linux-bionic-cuda11.6-py3.10-gcc7-sm86
+      docker-image: ${{ needs.linux-bionic-cuda11_6-py3_10-gcc7-sm86-build.outputs.docker-image }}
+      test-matrix: ${{ needs.linux-bionic-cuda11_6-py3_10-gcc7-sm86-build.outputs.test-matrix }}
+
   libtorch-linux-bionic-cuda11_7-py3_7-gcc7-build:
     name: libtorch-linux-bionic-cuda11.7-py3.7-gcc7
     uses: ./.github/workflows/_linux-build.yml