[quant][core][gpu][refactor] Refactored auxiliary functions in cudnn Conv.cpp to an utilities file (#73957)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/73957
Auxiliary functions in Conv.cpp that are applicable to a linear layer operator have been moved to
a utils.h file so that Conv.cpp and the future Linear.cpp file can both use it.
Test Plan:
This file is simply a refactorization and should only affect the cudnn
conv operator. We can run the following unit test:
```
python test/test_quantization.py TestQuantizedConv.test_qconv2d_cudnn
```
This file is simply a refactorization and should only affect the cudnn
conv operator. We can run the following unit test:
```
python test/test_quantization.py TestQuantizedConv.test_qconv2d_cudnn
```
Differential Revision:
D34995805
D34995805
Reviewed By: jerryzh168
Pulled By: dzdang
fbshipit-source-id: 15ea78af2927df583bf4adb596b255412186f902
(cherry picked from commit 6a1260c41868a5bb3721579e55ad00c738e8bd25)
