Implement more efficient pack and unpack uint5 #1138
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Summary: Implemented more efficient 5-bit packing and unpacking algorithm, for 8/64/128 values. The algorithm is commented in code, but you may also refer to T204077841 for discussion. ## Before ``` ---------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------- benchmark_pack_uint_values<1>/128/8 17.3 ns 17.2 ns 40530134 ... benchmark_pack_uint_values<5>/128/8 36.8 ns 36.5 ns 18974458 benchmark_pack_uint_values<5>/128/64 5.47 ns 5.43 ns 128341462 benchmark_pack_uint_values<5>/128/128 2.91 ns 2.70 ns 261633340 benchmark_unpack_uint_values<5>/128/8 28.8 ns 28.6 ns 24475696 benchmark_unpack_uint_values<5>/128/64 6.14 ns 5.65 ns 124953143 benchmark_unpack_uint_values<5>/128/128 2.90 ns 2.88 ns 242818639 ``` ## After ``` ---------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------- benchmark_pack_uint_values<1>/128/8 17.9 ns 17.5 ns 40221794 ... benchmark_pack_uint_values<5>/128/8 24.9 ns 24.8 ns 28330676 benchmark_pack_uint_values<5>/128/64 2.63 ns 2.61 ns 267856460 benchmark_pack_uint_values<5>/128/128 2.04 ns 2.03 ns 344166380 benchmark_unpack_uint_values<5>/128/8 22.1 ns 22.0 ns 31850032 benchmark_unpack_uint_values<5>/128/64 2.92 ns 2.89 ns 242508228 benchmark_unpack_uint_values<5>/128/128 2.33 ns 2.25 ns 310688575 ``` Reviewed By: metascroy Differential Revision: D64703548
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/1138
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 4071cc4 with merge base d84191c ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D64703548 |
metascroy
approved these changes
Oct 22, 2024
yanbing-j
pushed a commit
to yanbing-j/ao
that referenced
this pull request
Dec 9, 2024
Update AAR to 0919
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Implemented more efficient 5-bit packing and unpacking algorithm, for 8/64/128 values.
The algorithm is commented in code, but you may also refer to T204077841 for discussion.
Before
After
Reviewed By: metascroy
Differential Revision: D64703548