Add more eager global test

[wyg1997]

补全 eager global 单测：

• Sign
• SoftPlus （Fix softplus op #7888 先合并了，这个单测重复了）
• Slice

同时修复 0size slice op 后向崩溃的 bug

[github-actions]

Speed stats:

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/7818/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.4ms (= 12941.3ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.1ms (= 14209.1ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 142.1ms / 129.4ms)

OneFlow resnet50 time: 82.5ms (= 8249.5ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.9ms (= 8588.0ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.04 (= 85.9ms / 82.5ms)

OneFlow resnet50 time: 52.7ms (= 10533.9ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 56.3ms (= 11252.9ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.07 (= 56.3ms / 52.7ms)

OneFlow resnet50 time: 43.8ms (= 8764.8ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 46.2ms (= 9239.0ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.05 (= 46.2ms / 43.8ms)

OneFlow resnet50 time: 38.4ms (= 7688.8ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 34.2ms (= 6839.2ms / 200, input_shape=[1, 3, 224, 224])
❌ Relative speed: 0.89 (= 34.2ms / 38.4ms)

OneFlow swin dataloader time: 0.256s (= 51.113s / 200, num_workers=1)
PyTorch swin dataloader time: 0.150s (= 30.052s / 200, num_workers=1)
Relative speed: 0.588 (= 0.150s / 0.256s)

OneFlow swin dataloader time: 0.065s (= 12.981s / 200, num_workers=4)
PyTorch swin dataloader time: 0.042s (= 8.316s / 200, num_workers=4)
Relative speed: 0.641 (= 0.042s / 0.065s)

OneFlow swin dataloader time: 0.037s (= 7.446s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.500s / 200, num_workers=8)
Relative speed: 0.604 (= 0.023s / 0.037s)

❌ OneFlow resnet50 time: 147.1ms (= 14705.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 167.7ms (= 16766.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.14 (= 167.7ms / 147.1ms)

OneFlow resnet50 time: 100.6ms (= 10060.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 110.7ms (= 11071.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.10 (= 110.7ms / 100.6ms)

OneFlow resnet50 time: 74.6ms (= 14913.8ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 98.4ms (= 19685.0ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.32 (= 98.4ms / 74.6ms)

OneFlow resnet50 time: 65.0ms (= 12995.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.6ms (= 14923.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
❌ Relative speed: 1.15 (= 74.6ms / 65.0ms)

OneFlow resnet50 time: 57.2ms (= 11444.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.1ms (= 14816.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.29 (= 74.1ms / 57.2ms)

[github-actions]

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

[github-actions]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/7818/

[github-actions]

Speed stats:

GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.3ms (= 12931.8ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.6ms (= 14262.3ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 142.6ms / 129.3ms)

OneFlow resnet50 time: 80.3ms (= 8027.8ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 83.7ms (= 8371.6ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.04 (= 83.7ms / 80.3ms)

OneFlow resnet50 time: 54.3ms (= 10855.1ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 60.1ms (= 12010.7ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.11 (= 60.1ms / 54.3ms)

OneFlow resnet50 time: 41.2ms (= 8234.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 41.2ms (= 8231.3ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.00 (= 41.2ms / 41.2ms)

OneFlow resnet50 time: 36.6ms (= 7314.8ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 38.3ms (= 7655.9ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.05 (= 38.3ms / 36.6ms)

OneFlow swin dataloader time: 0.266s (= 53.234s / 200, num_workers=1)
PyTorch swin dataloader time: 0.152s (= 30.451s / 200, num_workers=1)
Relative speed: 0.572 (= 0.152s / 0.266s)

OneFlow swin dataloader time: 0.066s (= 13.151s / 200, num_workers=4)
PyTorch swin dataloader time: 0.042s (= 8.481s / 200, num_workers=4)
Relative speed: 0.645 (= 0.042s / 0.066s)

OneFlow swin dataloader time: 0.059s (= 11.876s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.562s / 200, num_workers=8)
Relative speed: 0.384 (= 0.023s / 0.059s)

❌ OneFlow resnet50 time: 146.6ms (= 14657.7ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 166.1ms (= 16614.1ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.13 (= 166.1ms / 146.6ms)

OneFlow resnet50 time: 96.0ms (= 9604.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 111.9ms (= 11194.2ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 111.9ms / 96.0ms)

OneFlow resnet50 time: 72.5ms (= 14509.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 93.2ms (= 18641.5ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.28 (= 93.2ms / 72.5ms)

OneFlow resnet50 time: 65.0ms (= 12993.5ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 80.7ms (= 16135.3ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.24 (= 80.7ms / 65.0ms)

OneFlow resnet50 time: 55.7ms (= 11130.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 67.6ms (= 13528.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.22 (= 67.6ms / 55.7ms)