ScalarLogicalKernel use primitive

[liujuncheng]

No description provided.

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8531/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.1ms (= 12909.1ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 143.6ms (= 14358.4ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.11 (= 143.6ms / 129.1ms)

OneFlow resnet50 time: 76.0ms (= 7599.4ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 86.5ms (= 8652.0ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.14 (= 86.5ms / 76.0ms)

OneFlow resnet50 time: 50.8ms (= 10156.9ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 56.4ms (= 11275.5ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.11 (= 56.4ms / 50.8ms)

OneFlow resnet50 time: 39.2ms (= 7846.3ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 46.6ms (= 9315.9ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.19 (= 46.6ms / 39.2ms)

OneFlow resnet50 time: 36.8ms (= 7350.5ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 36.8ms (= 7369.4ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.00 (= 36.8ms / 36.8ms)

OneFlow swin dataloader time: 0.266s (= 53.174s / 200, num_workers=1)
PyTorch swin dataloader time: 0.153s (= 30.663s / 200, num_workers=1)
Relative speed: 0.577 (= 0.153s / 0.266s)

OneFlow swin dataloader time: 0.074s (= 14.835s / 200, num_workers=4)
PyTorch swin dataloader time: 0.040s (= 8.022s / 200, num_workers=4)
Relative speed: 0.541 (= 0.040s / 0.074s)

OneFlow swin dataloader time: 0.040s (= 8.012s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.543s / 200, num_workers=8)
Relative speed: 0.567 (= 0.023s / 0.040s)

❌ OneFlow resnet50 time: 144.8ms (= 14481.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 169.8ms (= 16978.1ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 169.8ms / 144.8ms)

OneFlow resnet50 time: 93.0ms (= 9304.3ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 113.0ms (= 11297.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.21 (= 113.0ms / 93.0ms)

OneFlow resnet50 time: 70.0ms (= 13992.0ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 87.3ms (= 17468.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.25 (= 87.3ms / 70.0ms)

OneFlow resnet50 time: 57.2ms (= 11434.4ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 75.1ms (= 15022.3ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.31 (= 75.1ms / 57.2ms)

OneFlow resnet50 time: 52.4ms (= 10471.4ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 76.9ms (= 15389.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.47 (= 76.9ms / 52.4ms)