Fix compilation bug with CUDA 12.1 (#949)

* fix compilation
Signed-off-by: Edenzzzz <wtan45@wisc.edu>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Signed-off-by: Edenzzzz <wtan45@wisc.edu>

* Fix CUDA 12.1 compilation bug
Signed-off-by: Wenxuan(Eden) Tan <wtan45@wisc.edu>
Signed-off-by: Edenzzzz <wtan45@wisc.edu>

---------
Signed-off-by: Edenzzzz <wtan45@wisc.edu>
Signed-off-by: Wenxuan(Eden) Tan <wtan45@wisc.edu>
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

transformer_engine/pytorch/csrc/userbuffers/userbuffers.cu

@@ -5,6 +5,7 @@
  ************************************************************************/
 
 #include <cuda.h>
+#include <cuda_fp8.h>
 #include <cuda_runtime.h>
 
 #if __CUDA_ARCH__ >= 800
@@ -15,7 +16,6 @@
 #endif
 
 #include <assert.h>
-#include <cuda_fp8.h>
 #include <stdio.h>
 #include <unistd.h>