-
-
Notifications
You must be signed in to change notification settings - Fork 320
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
chore: update the performance tests to use @chainsafe/benchmark #7373
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## unstable #7373 +/- ##
============================================
+ Coverage 48.62% 50.22% +1.60%
============================================
Files 603 602 -1
Lines 40516 40407 -109
Branches 2071 2206 +135
============================================
+ Hits 19700 20294 +594
+ Misses 20778 20073 -705
- Partials 38 40 +2 |
c66f2f3
to
bcdad5f
Compare
|
||
// TODO: Diagnose why this benchmark failing after upgrade | ||
// https://github.com/ChainSafe/lodestar/issues/7380 | ||
bench.skip({ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Only this benchmark is failing, need to troubleshoot in particular later on.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just a couple small questions but looks good overall
@@ -38,7 +38,7 @@ describe("validate gossip attestation", () => { | |||
state, | |||
bitIndex: i, | |||
}); | |||
expect(subnet).to.be.equal(subnet0); | |||
assert.deepEqual(subnet, subnet0); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It seems weird that we need to use assert
here...
This reverts commit 8a0e51d.
It would be good to figure out #7380 before merging this |
there is no benchmark report / comment on this PR, this is no longer supported or just broken? |
Already looking into it. |
Performance Report✔️ no performance regression detected Full benchmark results
|
|
||
// TODO: Diagnose why this benchmark failing after upgrade | ||
// https://github.com/ChainSafe/lodestar/issues/7380 | ||
bench.skip({ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
would be good to see this addressed before merging
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will bring this back very soon on priority. Merging this PR now as needed it for another PR.
minRuns: 10 | ||
# Default is set to 0.005, which is too low considering the benchmark setup we have | ||
# Changing it to 0.05 which is 5/100, so 5% difference of moving average among run times | ||
convergeFactor: 0.075 # 7.5 / 100 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
how did you come up with this number? we should probably test if this is still low enough to catch actual regressions
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I run through some runs of few files and saw the outliers. Then approximates the outlier and came up with reasonable percentage to filter. And it's just an approximation, if the outliers skewed further the benchmark can fail.
Motivation
Use the
@chainsafe/benchmark
fork for our performance tests. This will enable to run these tests on multiple JS runtimes.Description
Steps to test or reproduce