Closed agnesLeroy closed 4 months ago
closes: please link all relevant issues
This PR fixes naming in gpu rotations and increases the number of threads in some kernels (to use the maximum possible number of threads always).
64 bit multi-bit benchmarks running on h100 -> https://github.com/zama-ai/tfhe-rs/actions/runs/9351572764
closes: please link all relevant issues
PR content/description
This PR fixes naming in gpu rotations and increases the number of threads in some kernels (to use the maximum possible number of threads always).
Check-list: