Use custom copy of cutlass to enable FP8 Grouped Gemm. (pytorch#726)

Summary:
Pull Request resolved: facebookresearch/FBGEMM#726

To support MOE models, we need to enable FP8 rowwise grouped gemm in FBGEMM. One missing piece to do this is support for rowwise scaling in cutlass. We have enabled this feature in a custom copy of cutlass but getting it into mainline will take a while. As a temporary measure, we can point FBGEMM to a custom copy of cutlass instead. Once the feature lands in mainline and we bump our support, we can go back to using the main repo.

X-link: pytorch#3649

Reviewed By: q10, jiawenliu64

Differential Revision: D68967944

Pulled By: jwfromm

fbshipit-source-id: 3e4625227ba6c33cf0478811fc9a8d40af361612

Author: jwfromm

.gitmodules

@@ -10,9 +10,12 @@
 [submodule "external/hipify_torch"]
 	path = external/hipify_torch
 	url = https://github.com/ROCmSoftwarePlatform/hipify_torch.git
+# TODO Using a private copy of cutlass is a temporary mitigation to enable grouped gemm.
+# Go back to main cutlass when possible.
 [submodule "external/cutlass"]
 	path = external/cutlass
-	url = https://github.com/NVIDIA/cutlass.git
+	url = https://github.com/jwfromm/cutlass.git
+	branch = FBGEMM
 [submodule "external/json"]
 	path = external/json
 	url = https://github.com/nlohmann/json.git