Add memset primitive

[liujuncheng]

No description provided.

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

OneFlow resnet50 time: 128.4ms (= 6418.7ms / 50, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 139.7ms (= 6985.6ms / 50, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.09 (= 139.7ms / 128.4ms)

OneFlow resnet50 time: 74.6ms (= 3729.0ms / 50, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.2ms (= 4258.4ms / 50, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.14 (= 85.2ms / 74.6ms)

OneFlow resnet50 time: 48.5ms (= 2426.3ms / 50, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.6ms (= 3082.3ms / 50, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.27 (= 61.6ms / 48.5ms)

OneFlow resnet50 time: 45.0ms (= 2250.7ms / 50, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 50.8ms (= 2541.7ms / 50, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.13 (= 50.8ms / 45.0ms)

OneFlow resnet50 time: 37.7ms (= 1883.2ms / 50, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 43.2ms (= 2159.7ms / 50, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.15 (= 43.2ms / 37.7ms)

OneFlow resnet50 time: 152.3ms (= 7615.8ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.4ms (= 8122.3ms / 50, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.07 (= 162.4ms / 152.3ms)

OneFlow resnet50 time: 97.3ms (= 4867.1ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 104.3ms (= 5216.1ms / 50, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.07 (= 104.3ms / 97.3ms)

OneFlow resnet50 time: 79.2ms (= 3962.2ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.8ms (= 3941.9ms / 50, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.99 (= 78.8ms / 79.2ms)

OneFlow resnet50 time: 73.6ms (= 3680.9ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 71.8ms (= 3588.3ms / 50, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.97 (= 71.8ms / 73.6ms)

OneFlow resnet50 time: 70.2ms (= 3511.5ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 59.5ms (= 2972.8ms / 50, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 0.85 (= 59.5ms / 70.2ms)