Switch back to NVIDIA/cutlass, and upgrade to v4.2.1

Author: jasl

.gitmodules

@@ -10,12 +10,9 @@
 [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/jwfromm/cutlass
-	branch = FBGEMM
+	url = https://github.com/NVIDIA/cutlass.git
 [submodule "external/composable_kernel"]
 	path = external/composable_kernel
 	url = https://github.com/ROCm/composable_kernel.git