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

[WIP] Move elementwise based kernels to use peeling approach. #9564

Closed
wants to merge 1 commit into from
Closed

[WIP] Move elementwise based kernels to use peeling approach. #9564

wants to merge 1 commit into from

Conversation

hanhanW
Copy link
Contributor

@hanhanW hanhanW commented Jun 21, 2022

No description provided.

@hanhanW hanhanW requested a review from dcaballe June 21, 2022 08:26
@hanhanW hanhanW mentioned this pull request Jun 21, 2022
Closed
@iree-github-actions-bot
Copy link
Contributor

iree-github-actions-bot commented Jun 21, 2022 •
edited
Loading

Abbreviated x86_64 Benchmark Summary (experimental)

@ commit b4a466bffd4d0d654c8886eb896b496aea704173 (vs. base 494fa16143c60de167f54b3c63d2ac98675d4d76)

No improved or regressed benchmarks 🏖️

For more information:

@iree-github-actions-bot
Copy link
Contributor

iree-github-actions-bot commented Jun 21, 2022 •
edited
Loading

Abbreviated Benchmark Summary

@ commit b4a466bffd4d0d654c8886eb896b496aea704173 (vs. base 494fa16143c60de167f54b3c63d2ac98675d4d76)

Regressed Benchmarks 🚩

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
DeepLabV3 [fp32] (TFLite) 4-thread,little-core,full-inference,default-flags with IREE-Dylib @ Pixel-6-Pro (CPU-ARMv8.2-A) 153 (vs. 131, 16.79%↑) 153 3
MobileBertSquad [int8] (TFLite) 4-thread,big-core,full-inference,experimental-flags with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 232 (vs. 203, 14.29%↑) 243 19
MobileSSD [fp32] (TFLite) big-core,full-inference,experimental-flags with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 98 (vs. 90, 8.89%↑) 98 0

[Top 3 out of 5 results showed]

Improved Benchmarks 🎉

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
MobileNetV3Small [fp32,imagenet] (TFLite) little-core,full-inference,experimental-flags with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 56 (vs. 63, 11.11%↓) 56 0
MobileNetV3Small [fp32,imagenet] (TFLite) 4-thread,big-core,full-inference,experimental-flags with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 10 (vs. 11, 9.09%↓) 10 0
MobileBertSquad [fp32] (TFLite) 4-thread,big-core,full-inference,experimental-flags with IREE-Dylib @ Pixel-6-Pro (CPU-ARMv8.2-A) 414 (vs. 455, 9.01%↓) 436 49

[Top 3 out of 9 results showed]

For more information:

@hanhanW hanhanW mentioned this pull request Jun 29, 2022
Closed
@benvanik benvanik changed the title [WIP] Move elemnwise based kernels to use peeling approach. [WIP] Move elementwise based kernels to use peeling approach. Jul 1, 2022
@hanhanW hanhanW closed this Jul 18, 2022
@hanhanW hanhanW deleted the cpu-elem-peeling branch July 18, 2022 05:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@dcaballe dcaballe Awaiting requested review from dcaballe

@MaheshRavishankar MaheshRavishankar Awaiting requested review from MaheshRavishankar MaheshRavishankar will be requested when the pull request is marked ready for review MaheshRavishankar is a code owner

Assignees
No one assigned
Labels
(deprecated) buildkite:benchmark-android Deprecated. Please use benchmarks:android-*
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@hanhanW @iree-github-actions-bot