Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Update llvm version and refine IREE cases #8461

Merged
merged 5 commits into from
Jun 23, 2022
Merged

Update llvm version and refine IREE cases #8461

merged 5 commits into from
Jun 23, 2022

Conversation

jackalcooper
Copy link
Collaborator

No description provided.

@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8461/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.7ms (= 12973.0ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.8ms (= 14279.9ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 142.8ms / 129.7ms)

OneFlow resnet50 time: 80.0ms (= 8002.1ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.3ms (= 8532.9ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.07 (= 85.3ms / 80.0ms)

OneFlow resnet50 time: 49.6ms (= 9927.9ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 58.0ms (= 11608.3ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.17 (= 58.0ms / 49.6ms)

OneFlow resnet50 time: 43.2ms (= 8642.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.0ms (= 8797.5ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.02 (= 44.0ms / 43.2ms)

OneFlow resnet50 time: 38.3ms (= 7652.5ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 42.7ms (= 8547.2ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.12 (= 42.7ms / 38.3ms)

OneFlow swin dataloader time: 0.245s (= 49.087s / 200, num_workers=1)
PyTorch swin dataloader time: 0.152s (= 30.368s / 200, num_workers=1)
Relative speed: 0.619 (= 0.152s / 0.245s)

OneFlow swin dataloader time: 0.065s (= 13.012s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.241s / 200, num_workers=4)
Relative speed: 0.633 (= 0.041s / 0.065s)

OneFlow swin dataloader time: 0.037s (= 7.305s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.391s / 200, num_workers=8)
Relative speed: 0.601 (= 0.022s / 0.037s)

❌ OneFlow resnet50 time: 146.2ms (= 14621.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 166.7ms (= 16665.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.14 (= 166.7ms / 146.2ms)

OneFlow resnet50 time: 95.2ms (= 9516.1ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 113.2ms (= 11317.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 113.2ms / 95.2ms)

OneFlow resnet50 time: 71.3ms (= 14269.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 89.1ms (= 17822.5ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.25 (= 89.1ms / 71.3ms)

OneFlow resnet50 time: 62.0ms (= 12401.6ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 76.9ms (= 15384.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.24 (= 76.9ms / 62.0ms)

OneFlow resnet50 time: 55.0ms (= 11008.4ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 69.2ms (= 13844.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.26 (= 69.2ms / 55.0ms)

@github-actions
Copy link
Contributor

Static analysis with clang failed. PR label automerge has been removed

@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8461/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.7ms (= 12973.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 145.0ms (= 14496.8ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.12 (= 145.0ms / 129.7ms)

OneFlow resnet50 time: 76.1ms (= 7609.4ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 84.7ms (= 8470.4ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.11 (= 84.7ms / 76.1ms)

OneFlow resnet50 time: 50.7ms (= 10134.1ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 58.9ms (= 11771.4ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.16 (= 58.9ms / 50.7ms)

OneFlow resnet50 time: 42.7ms (= 8532.6ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 41.5ms (= 8302.7ms / 200, input_shape=[2, 3, 224, 224])
❌ Relative speed: 0.97 (= 41.5ms / 42.7ms)

OneFlow resnet50 time: 37.9ms (= 7571.7ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 45.6ms (= 9120.7ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.20 (= 45.6ms / 37.9ms)

OneFlow swin dataloader time: 0.383s (= 76.602s / 200, num_workers=1)
PyTorch swin dataloader time: 0.152s (= 30.428s / 200, num_workers=1)
Relative speed: 0.397 (= 0.152s / 0.383s)

OneFlow swin dataloader time: 0.064s (= 12.850s / 200, num_workers=4)
PyTorch swin dataloader time: 0.041s (= 8.251s / 200, num_workers=4)
Relative speed: 0.642 (= 0.041s / 0.064s)

OneFlow swin dataloader time: 0.037s (= 7.393s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.449s / 200, num_workers=8)
Relative speed: 0.602 (= 0.022s / 0.037s)

❌ OneFlow resnet50 time: 145.9ms (= 14591.8ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 167.6ms (= 16761.5ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.15 (= 167.6ms / 145.9ms)

OneFlow resnet50 time: 96.0ms (= 9599.4ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 113.5ms (= 11349.5ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.18 (= 113.5ms / 96.0ms)

OneFlow resnet50 time: 71.3ms (= 14252.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 87.0ms (= 17406.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.22 (= 87.0ms / 71.3ms)

OneFlow resnet50 time: 60.9ms (= 12177.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.8ms (= 14957.1ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.23 (= 74.8ms / 60.9ms)

OneFlow resnet50 time: 54.6ms (= 10927.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 66.4ms (= 13287.8ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.22 (= 66.4ms / 54.6ms)

@github-actions
Copy link
Contributor

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

@jackalcooper jackalcooper enabled auto-merge (squash) June 23, 2022 22:05
@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8461/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.8ms (= 12980.4ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 143.4ms (= 14336.1ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.10 (= 143.4ms / 129.8ms)

OneFlow resnet50 time: 76.0ms (= 7596.5ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 87.8ms (= 8779.6ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.16 (= 87.8ms / 76.0ms)

OneFlow resnet50 time: 48.5ms (= 9707.5ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.9ms (= 12383.2ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.28 (= 61.9ms / 48.5ms)

OneFlow resnet50 time: 39.5ms (= 7900.9ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 44.0ms (= 8798.3ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.11 (= 44.0ms / 39.5ms)

OneFlow resnet50 time: 35.1ms (= 7011.5ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 42.1ms (= 8424.2ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.20 (= 42.1ms / 35.1ms)

OneFlow swin dataloader time: 0.274s (= 54.898s / 200, num_workers=1)
PyTorch swin dataloader time: 0.151s (= 30.219s / 200, num_workers=1)
Relative speed: 0.550 (= 0.151s / 0.274s)

OneFlow swin dataloader time: 0.107s (= 21.496s / 200, num_workers=4)
PyTorch swin dataloader time: 0.040s (= 8.045s / 200, num_workers=4)
Relative speed: 0.374 (= 0.040s / 0.107s)

OneFlow swin dataloader time: 0.040s (= 7.984s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.374s / 200, num_workers=8)
Relative speed: 0.548 (= 0.022s / 0.040s)

❌ OneFlow resnet50 time: 146.3ms (= 14626.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 169.4ms (= 16937.7ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 169.4ms / 146.3ms)

OneFlow resnet50 time: 94.4ms (= 9441.6ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 112.5ms (= 11252.8ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 112.5ms / 94.4ms)

OneFlow resnet50 time: 71.5ms (= 14290.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 98.4ms (= 19682.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.38 (= 98.4ms / 71.5ms)

OneFlow resnet50 time: 56.2ms (= 11249.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.0ms (= 14808.2ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.32 (= 74.0ms / 56.2ms)

OneFlow resnet50 time: 49.0ms (= 9801.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 78.4ms (= 15677.1ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.60 (= 78.4ms / 49.0ms)

@jackalcooper jackalcooper merged commit cd66d3d into master Jun 23, 2022
@jackalcooper jackalcooper deleted the upgrade-llvm-and-refine-lit-cases branch June 23, 2022 23:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@mosout mosout mosout approved these changes

@BBuf BBuf BBuf approved these changes

@howin98 howin98 howin98 approved these changes

@oneflow-ci-bot oneflow-ci-bot Awaiting requested review from oneflow-ci-bot

@daquexian daquexian Awaiting requested review from daquexian

@hjchen2 hjchen2 Awaiting requested review from hjchen2 hjchen2 is a code owner

Assignees
No one assigned
Labels
automerge ci enhancement ir
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
4 participants
@jackalcooper @mosout @BBuf @howin98