-
Notifications
You must be signed in to change notification settings - Fork 29.8k
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
test,crypto: sequential/test-crypto-timing-safe-equal-benchmarks flaky locally #8744
Comments
How often does this happen, roughly? Is it always that error ( |
Reasonably often. It's quite reliably fails on one machine:
The other test seems solid:
|
Within It looks like the |
I can't seem to reproduce the issue on OS X El Capitan 10.11.6, on the current version of master (782620f). Is the issue happening for you on Linux? edit: Also, is the issue happening on any of the CI machines? If so, that will make it easier to debug. |
Yes, this is on an otherwise unremarkable x86_64 FC24 laptop. I don't have an OS X machine at hand at the moment. I haven't (personally) seen the test fail on the CI recently. |
Still fails for me every now and then on that one machine. Just now:
|
Could you try increasing this number by an order of magnitude or so, to see if that fixes the issue on that machine? |
With numTrials = 1e5, the test passes five times in a row:
Aside: 6 seconds for a sequential test is really too long. Sequential tests should ideally finish in under a second. |
Where should the test go instead? Pummel? This is sort of an unusual test in that it requires running a very large number of benchmarks, so reducing the runtime without making it flakier is probably going to be non-trivial. But I'd be happy to move it to a different place if that makes it less annoying while developing. |
Pummel is probably the best place for it. The one downside is that pummel tests aren't run regularly. |
fyi... it just failed here also: https://ci.nodejs.org/job/node-test-commit-freebsd/4945/nodes=freebsd10-64/console ... this flakiness seems relatively new. |
Increase the number of iterations from 1e4 to 1e5. Makes the test pass for me locally when previously it would fail 9 out of 10 times because the running time was not enough to smooth away the outliers. Fixes: nodejs#8744 PR-URL: nodejs#9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Move sequential/test-crypto-timing-safe-equal-benchmarks to test/pummel because it fails for me locally quite frequently and because it takes about five or six seconds to complete, which is too long for a test in test/sequential. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Increase the number of iterations from 1e4 to 1e5. Makes the test pass for me locally when previously it would fail 9 out of 10 times because the running time was not enough to smooth away the outliers. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Move sequential/test-crypto-timing-safe-equal-benchmarks to test/pummel because it fails for me locally quite frequently and because it takes about five or six seconds to complete, which is too long for a test in test/sequential. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Increase the number of iterations from 1e4 to 1e5. Makes the test pass for me locally when previously it would fail 9 out of 10 times because the running time was not enough to smooth away the outliers. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Move sequential/test-crypto-timing-safe-equal-benchmarks to test/pummel because it fails for me locally quite frequently and because it takes about five or six seconds to complete, which is too long for a test in test/sequential. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Increase the number of iterations from 1e4 to 1e5. Makes the test pass for me locally when previously it would fail 9 out of 10 times because the running time was not enough to smooth away the outliers. Fixes: #8744 PR-URL: #9241 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: not-an-aardvark <[email protected]>
Running
make test
on master I occasionally get this:cc @not-an-aardvark
The text was updated successfully, but these errors were encountered: