eager global masked select

[liufengwei0103]

fix https://github.com/Oneflow-Inc/OneTeam/issues/1055

[github-actions]

Code got formatted by CI. Please request CI again if you still want to have this PR merged. If the PR is from a forked repo, please download the patch files from the GitHub Actions web page and apply them locally.

[github-actions]

CI failed when running job: cuda-module-distributed. PR label automerge has been removed

[github-actions]

CI failed when running job: cuda-module-distributed-rank-1. PR label automerge has been removed

[github-actions]

CI failed when running job: cuda-module-distributed-rank-0. PR label automerge has been removed

[github-actions]

CI failed when running job: cuda-module-distributed-rank-1. PR label automerge has been removed

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.7ms (= 12873.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 137.2ms (= 13715.4ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.07 (= 137.2ms / 128.7ms)

✔️ OneFlow resnet50 time: 78.4ms (= 7842.7ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 83.7ms (= 8369.8ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.07 (= 83.7ms / 78.4ms)

OneFlow resnet50 time: 52.9ms (= 10577.1ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 60.4ms (= 12071.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.14 (= 60.4ms / 52.9ms)

OneFlow resnet50 time: 41.0ms (= 8198.6ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 47.7ms (= 9535.0ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.16 (= 47.7ms / 41.0ms)

OneFlow resnet50 time: 38.2ms (= 7636.8ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 43.9ms (= 8786.4ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.15 (= 43.9ms / 38.2ms)

✔️ OneFlow resnet50 time: 142.7ms (= 14269.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 160.1ms (= 16009.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.12 (= 160.1ms / 142.7ms)

OneFlow resnet50 time: 90.0ms (= 9003.2ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 103.6ms (= 10358.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.15 (= 103.6ms / 90.0ms)

OneFlow resnet50 time: 62.4ms (= 12471.8ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 75.6ms (= 15117.0ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.21 (= 75.6ms / 62.4ms)

OneFlow resnet50 time: 53.2ms (= 10634.4ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 67.6ms (= 13525.6ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.27 (= 67.6ms / 53.2ms)

OneFlow resnet50 time: 47.6ms (= 9514.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 64.4ms (= 12887.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.35 (= 64.4ms / 47.6ms)