onnxruntime

QuickGelu Fusion (#12417)

Some models have QuickGelu(x)=x*sigmoid(1.702x), which has 3 Ops for
forward and 5 Ops for backward. The PR is to fuse this to a single Op
named QuickGelu and its gradient QuickGeluGrad.

For CUDA, tested in V100 using input tensor with shape [64,128,2048] and
float16 type:
Before, FW takes 335us, BW takes 614us

![image](https://user-images.githubusercontent.com/11661208/182291335-15188709-ffe7-44d1-9d14-0b544cbe5e55.png)

After, FW takes 115us, BW takes 139us, which is much faster.

![image](https://user-images.githubusercontent.com/11661208/182291502-f0b5161c-b95c-45fc-90f8-ad0c592d2433.png)

For CPU kernel, using same shape and float type:
Before, FW takes 10us, BW takes 49us
Mul: 3480[µs]
Sigmoid: 1996[µs]
Mul: 4789[µs]
Mul: 4642[µs]
Mul: 4195[µs]
SigmoidGrad: 18328[µs]
Mul: 2988[µs]
Sum: 18576[µs]

After, FW takes 4us, BW takes 5us, which is also much faster.
QuickGelu: 3939[µs]
QuickGeluGrad: 5089[µs]

Co-authored-by: Vincent Wang <weicwang@microsoft.com>