-
Notifications
You must be signed in to change notification settings - Fork 302
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.
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
Update gloo benchmark output #294
Closed
Closed
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
This pull request was exported from Phabricator. Differential Revision: D26637016 |
sofong5
pushed a commit
to sofong5/gloo
that referenced
this pull request
Mar 2, 2021
Summary: Pull Request resolved: facebookincubator#294 Design Doc: https://fb.quip.com/vWFvAIF4HK0f The purpose of this diff is to update the output format of the gloo benchmark. More specifically, we want to: - Update how errors are displayed during correctness verification. - Add sections for better readability - Clarify column titles For mismatch errors: Originally, the benchmark would exit right away once it detects a mismatch. This diff changes this so that the currently running benchmark is now run to completion and mismatches are only displayed at the very end. If a user does not specify element size, the benchmark by default will do a sweep of numbers ranging from 100 to 5000000. In this case, the benchmark will stop the next sweep at the first instance of a mismatch (but will still run current one to completion). See test plan for example. Change Summary: - Print value of verify flag in options - Refactor out a common verify pattern into a new function `constStrideVerify` where there is a constant stride between elements in the array - Add new `errors` parameter to the verify function which will catch all error messages generated to print at the end - New class function `checkErrors` that prints the errors and exits - New class functions `printVerifyHeader` and `printFooter` - Break output into more readable sections by printing headers - Add new `size` column to benchmark results - Set default for verify to be `true` - Add new option `--show-all-errors` to display all errors (default will truncate to 20 errors if number of errors is greater or equal to 100) Differential Revision: D26637016 fbshipit-source-id: 5589760c3325e29f2828fe2e75390d4b73f6b2ca
This pull request was exported from Phabricator. Differential Revision: D26637016 |
sofong5
force-pushed
the
export-D26637016
branch
from
March 2, 2021 22:01
fca3d82
to
5b22517
Compare
sofong5
changed the title
Update how mismatch errors are displayed during verify
Update gloo benchmark output
Mar 2, 2021
Summary: Pull Request resolved: facebookincubator#294 Design Doc: https://fb.quip.com/vWFvAIF4HK0f The purpose of this diff is to update the output format of the gloo benchmark. More specifically, we want to: - Update how errors are displayed during correctness verification. - Add sections for better readability - Clarify column titles For mismatch errors: Originally, the benchmark would exit right away once it detects a mismatch. This diff changes this so that the currently running benchmark is now run to completion and mismatches are only displayed at the very end. If a user does not specify element size, the benchmark by default will do a sweep of numbers ranging from 100 to 5000000. In this case, the benchmark will stop the next sweep at the first instance of a mismatch (but will still run current one to completion). See test plan for example. Change Summary: - Print value of verify flag in options - Refactor out a common verify pattern into a new function `constStrideVerify` where there is a constant stride between elements in the array - Add new `errors` parameter to the verify function which will catch all error messages generated to print at the end - New class function `checkErrors` that prints the errors and exits - New class functions `printVerifyHeader` and `printFooter` - Break output into more readable sections by printing headers - Add new `size` column to benchmark results - Set default for verify to be `true` - Add new option `--show-all-errors` to display all errors (default will truncate to 20 errors if number of errors is greater or equal to 100) Reviewed By: osalpekar Differential Revision: D26637016 fbshipit-source-id: 9dc2a20984861b9e9e30c3bdf657174797fd8b95
sofong5
force-pushed
the
export-D26637016
branch
from
March 5, 2021 00:54
5b22517
to
a7cea6a
Compare
This pull request was exported from Phabricator. Differential Revision: D26637016 |
This pull request has been merged in 4b41412. |
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:
The purpose of this diff is to update how errors are displayed during correctness verification.
Originally, the benchmark would exit right away once it detects a mismatch. This diff changes this so that the currently running benchmark is now run to completion and mismatches are only displayed at the very end.
If a user does not specify element size, the benchmark by default will do a sweep of numbers ranging from 100 to 5000000. In this case, the benchmark will stop the next sweep at the first instance of a mismatch (but will still run current one to completion). See test plan for example.
Change Summary:
constStrideVerify
where there is a constant stride between elements in the arrayprintMismatches
that outputs the mismatchesDifferential Revision: D26637016