eager global ne

[liufengwei0103]

No description provided.

[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 (= 12874.2ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 140.5ms (= 14048.0ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.09 (= 140.5ms / 128.7ms)

❌ OneFlow resnet50 time: 79.0ms (= 7898.3ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 88.4ms (= 8844.6ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.12 (= 88.4ms / 79.0ms)

OneFlow resnet50 time: 53.8ms (= 10754.1ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 55.2ms (= 11043.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.03 (= 55.2ms / 53.8ms)

OneFlow resnet50 time: 42.8ms (= 8565.8ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 47.8ms (= 9556.4ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.12 (= 47.8ms / 42.8ms)

OneFlow resnet50 time: 35.5ms (= 7096.6ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 40.3ms (= 8059.2ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.14 (= 40.3ms / 35.5ms)

✔️ OneFlow resnet50 time: 142.3ms (= 14228.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 163.9ms (= 16389.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.15 (= 163.9ms / 142.3ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 89.6ms (= 8960.2ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 104.5ms (= 10446.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 104.5ms / 89.6ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 62.5ms (= 12493.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 76.4ms (= 15285.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.22 (= 76.4ms / 62.5ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 53.8ms (= 10756.1ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 67.4ms (= 13478.6ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.25 (= 67.4ms / 53.8ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 47.2ms (= 9447.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 62.3ms (= 12450.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.32 (= 62.3ms / 47.2ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

[github-actions]

CI failed when running job: cuda-speed-test. PR label automerge has been removed

[liufengwei0103]

test case不影响速度，重跑

[github-actions]

Speed stats:

GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.6ms (= 12862.8ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 137.1ms (= 13706.2ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.07 (= 137.1ms / 128.6ms)

✔️ OneFlow resnet50 time: 77.4ms (= 7742.5ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 87.2ms (= 8716.4ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.13 (= 87.2ms / 77.4ms)

OneFlow resnet50 time: 52.8ms (= 10551.6ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 62.0ms (= 12397.7ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.17 (= 62.0ms / 52.8ms)

OneFlow resnet50 time: 44.0ms (= 8808.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 51.4ms (= 10280.2ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.17 (= 51.4ms / 44.0ms)

OneFlow resnet50 time: 39.0ms (= 7803.9ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 40.1ms (= 8011.5ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.03 (= 40.1ms / 39.0ms)

✔️ OneFlow resnet50 time: 142.6ms (= 14261.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 156.1ms (= 15611.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.09 (= 156.1ms / 142.6ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 86.2ms (= 8618.1ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 103.6ms (= 10360.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.20 (= 103.6ms / 86.2ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 62.5ms (= 12509.4ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 77.0ms (= 15407.0ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.23 (= 77.0ms / 62.5ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 50.7ms (= 10130.5ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 66.9ms (= 13383.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.32 (= 66.9ms / 50.7ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************

OneFlow resnet50 time: 49.7ms (= 9945.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.7ms (= 13739.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.38 (= 68.7ms / 49.7ms)
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************