Skip to content
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

Singlebranch #5251

Closed
Closed

Conversation

magnumripper
Copy link
Member

@magnumripper magnumripper commented Mar 13, 2023

Third try.

Now there's just a single if unlikely() in the end of the function, but AMD still acts up with poor performance 😢

@magnumripper magnumripper marked this pull request as draft March 13, 2023 00:49
@magnumripper magnumripper mentioned this pull request Mar 13, 2023
@magnumripper
Copy link
Member Author

magnumripper commented Mar 13, 2023

Now there's just a single if unlikely() in the end of the function, but AMD still acts up with poor performance 😢

That was a commit mistake. Performance is up a good deal now, 19144M seen. Lowering from 125 to 59 doesn't make any difference.

@magnumripper magnumripper force-pushed the singlebranch branch 7 times, most recently from 2cac6d5 to 7b132a4 Compare March 14, 2023 22:15
Found when implementing a 64-bit version of it.  The impact would be
suboptimal bitmap parameters when adjusting bitmap as the number of
remaining hashes decreases.
Not only do we save memory, we can reverse much more as well!  We check
the remaining bits in cold host code, for good measure.
This way we still got optimal speed for the original nt-opencl format that
support lengths up to 27.

1 block of MD4 is up to 27 characters.
2 blocks is 59, 3 is 91, 4 is 123 and 5 is 125 (due to core max).

As long as we bump it over 27 we don't seem to gain any speed by limiting
it to less than 125 characters.

See openwall#5245
@magnumripper magnumripper deleted the singlebranch branch March 15, 2023 18:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant