Add sync data for to_global function

[hanwen-sun]

为to_global function添加sync_data参数，在python端调用时可指定参数sync_data为true or false, 默认值为true。相应的, local_to_global function添加sync_data参数。修改对应的docstr文档。

[github-actions[bot]]

View latest API docs preview at: https://oneflow-staging.oss-cn-beijing.aliyuncs.com/docs/Oneflow-Inc/oneflow/pr/10433/

[github-actions[bot]]

Speed stats:

GPU Name: NVIDIA GeForce RTX 3080 Ti 

❌ OneFlow resnet50 time: 43.6ms (= 4359.1ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 57.8ms (= 5776.1ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.33 (= 57.8ms / 43.6ms)

OneFlow resnet50 time: 26.6ms (= 2655.9ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 37.5ms (= 3753.8ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.41 (= 37.5ms / 26.6ms)

OneFlow resnet50 time: 19.0ms (= 3790.2ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 36.2ms (= 7235.1ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.91 (= 36.2ms / 19.0ms)

OneFlow resnet50 time: 17.8ms (= 3561.8ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 31.7ms (= 6330.7ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.78 (= 31.7ms / 17.8ms)

OneFlow resnet50 time: 17.3ms (= 3451.9ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 29.0ms (= 5792.2ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.68 (= 29.0ms / 17.3ms)

OneFlow swin dataloader time: 0.200s (= 40.028s / 200, num_workers=1)
PyTorch swin dataloader time: 0.128s (= 25.597s / 200, num_workers=1)
Relative speed: 0.639 (= 0.128s / 0.200s)

OneFlow swin dataloader time: 0.059s (= 11.782s / 200, num_workers=4)
PyTorch swin dataloader time: 0.032s (= 6.441s / 200, num_workers=4)
Relative speed: 0.547 (= 0.032s / 0.059s)

OneFlow swin dataloader time: 0.032s (= 6.302s / 200, num_workers=8)
PyTorch swin dataloader time: 0.017s (= 3.315s / 200, num_workers=8)
Relative speed: 0.526 (= 0.017s / 0.032s)

❌ OneFlow resnet50 time: 49.1ms (= 4909.8ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 65.8ms (= 6579.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.34 (= 65.8ms / 49.1ms)

OneFlow resnet50 time: 36.7ms (= 3672.3ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 46.7ms (= 4672.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.27 (= 46.7ms / 36.7ms)

OneFlow resnet50 time: 28.3ms (= 5651.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 39.8ms (= 7950.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.41 (= 39.8ms / 28.3ms)

OneFlow resnet50 time: 25.0ms (= 4998.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 38.7ms (= 7744.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.55 (= 38.7ms / 25.0ms)

OneFlow resnet50 time: 24.0ms (= 4792.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 36.1ms (= 7222.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.51 (= 36.1ms / 24.0ms)