🐛 Bug

Compile Failure with cuda 92

To Reproduce

Steps to reproduce the behavior:

Compile PyTorch with CUDA92 USE_CUDA=1 USE_NCCL=0 USE_DISTRIBUTED=0 python setup.py develop

/home/erjia/workspace/pytorch_dev/typing/c10/util/Optional.h(641): error: identifier "setInitialized" is undefined

1 error detected in the compilation of "/tmp/tmpxft_000705f0_00000000-6_THCTensorMathPointwiseHalf.cpp1.ii".
CMake Error at torch_cuda_generated_THCTensorMathPointwiseHalf.cu.o.Release.cmake:281 (message):
    Error generating file
    /home/erjia/workspace/pytorch_dev/typing/build/caffe2/CMakeFiles/torch_cuda.dir/__/aten/src/THC/generated/./torch_cuda_generated_THCTensorMathPointwiseHalf.cu.o

Expected behavior

No Error

Environment

Collecting environment information...
PyTorch version: 1.10.0a0+git1fc3576
Is debug build: False
CUDA used to build PyTorch: 9.2
ROCM used to build PyTorch: N/A

OS: CentOS Linux 7 (Core) (x86_64)
GCC version: (GCC) 7.3.1 20180303 (Red Hat 7.3.1-5)
Clang version: Could not collect
CMake version: version 3.19.6
Libc version: glibc-2.17

Python version: 3.9 (64-bit runtime)
Python platform: Linux-5.2.9-229_fbk15_hardened_4185_g357f49b36602-x86_64-with-glibc2.17
Is CUDA available: True
CUDA runtime version: 9.2.88
GPU models and configuration:
GPU 0: Tesla M40
GPU 1: Tesla M40

Nvidia driver version: 450.80.02
cuDNN version: /usr/local/cuda-9.2/targets/x86_64-linux/lib/libcudnn.so.7
HIP runtime version: N/A
MIOpen runtime version: N/A

Additional context

Possibly due to the change in the PR (#59333)

Fix in #60017

cc @malfet @seemethere @walterddr @ngimel