Skip to content

Conversation

dop-amin
Copy link
Collaborator

@dop-amin dop-amin commented May 23, 2025

Adds a simple framework for running regression tetsts on SLOTHY's performance to CI.
Partially addresses #215.
Depends on #211.

@dop-amin dop-amin marked this pull request as ready for review May 23, 2025 15:36
@dop-amin dop-amin assigned dop-amin and unassigned dop-amin May 23, 2025
@dop-amin dop-amin marked this pull request as draft May 23, 2025 15:42
@dop-amin dop-amin force-pushed the regression branch 3 times, most recently from c587fba to dc9f5ea Compare May 27, 2025 14:32
@dop-amin
Copy link
Collaborator Author

For some reason ntt_dilithium_123_45678_a55 keeps failing due to immediate out of range issues. This was not the case with the previous optimization run in CI. Needs investigation.

@mkannwischer mkannwischer force-pushed the regression branch 2 times, most recently from b4ed8cb to 37e29ee Compare June 1, 2025 08:26
@mkannwischer
Copy link
Collaborator

Here is the first test result:

keccak_f1600_x1_scalar_slothy_no_symbolic_a55: 2325.6789305686953
ntt_kyber_123_4567_a55: 249.27907199859618
keccakf1600_xkcp_m7: 1728.8838932037354
ntt_dilithium_m7: 16.455983209609986
ntt_kyber_m7: 2825.1233290195464
ntt_kyber_1_23_45_67_m55: 27.949791622161865
ntt_dilithium_12_34_56_78_m55: 28.257723569869995

I'll run it again to see how stable it is.

@dop-amin
Copy link
Collaborator Author

Could also be worth returning the variance

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.

2 participants