-
Notifications
You must be signed in to change notification settings - Fork 2k
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
Fix the connectblockslow
testvector generator script to be deterministic.
#2388
Labels
A-CI
Area: Continuous Integration
A-testing
Area: Tests and testing infrastructure
I-performance
Problems and improvements with respect to performance
M-has-pr
To-be-removed (GitHub has linked:pr filter)
Projects
Comments
@bitcartel sent me the archive he made, and the non-determinism is obvious:
The owner and group can easily be made deterministic with that |
Looked at how the Gitian descriptors do it - |
str4d
added a commit
to str4d/zcash
that referenced
this issue
May 20, 2017
The archive has also been moved from .tar.gz to .tar.xz for a 33% reduction in size. Closes zcash#2388.
zkbot
added a commit
that referenced
this issue
May 22, 2017
…at-least Remove additional sources of nondeterminism from benchmark archive Closes #2388.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
A-CI
Area: Continuous Integration
A-testing
Area: Tests and testing infrastructure
I-performance
Problems and improvements with respect to performance
M-has-pr
To-be-removed (GitHub has linked:pr filter)
In this PR review @bitcartel determined that he could not reproduce the same hash for the
connectblockslow
testvector that @str4d created.We timeboxed fixing that non-determinism and hit the time limit, and decided to deploy with a 'baked in' non-deterministic testvector for the time being.
The purpose of this ticket is to fix that non-determinism so that anyone can recreate the same benchmark on identical data. After that is done, we should measure that against today's 'special instance' of the test vector to ensure there's no performance difference, then update our benchmarking system to use the deterministic dataset.
This is important for a variety of reasons. One is to detect regressions in the determinism itself, as well as to ensure repeatability of measurements.
The text was updated successfully, but these errors were encountered: