refactor(gpu): refactor sample extract and modulus switch to match CPU's version

zama-ai / tfhe-rs

TFHE-rs: A Pure Rust implementation of the TFHE Scheme for Boolean and Integer Arithmetics Over Encrypted Data.

Other

907 stars 143 forks source link

refactor(gpu): refactor sample extract and modulus switch to match CPU's version #1414

Closed pdroalves closed 2 months ago

pdroalves commented 2 months ago

…

closes: https://github.com/zama-ai/tfhe-rs-internal/issues/668

PR content/description

Check-list:

[ ] Tests for the changes have been added (for bug fixes / features)
[ ] Docs have been added / updated (for bug fixes / features)
[ ] Relevant issues are marked as resolved/closed, related issues are linked in the description
[ ] Check for breaking changes (including serialization changes) and add them to commit message following the conventional commit specification

agnesLeroy commented 2 months ago

@pdroalves the 64-bit benchmark results on H100 are the same for all operations except mul/shift/rotate, which are better: Screenshot from 2024-07-25 11-38-20