juuso-oskari 95f813013f Add compile-time MaxNumBlocks optimization ...

- Added MaxNumBlocks template parameter to all kernel traits
- Propagated through pipeline problem and pipeline
- Added compile-time kNeedsRebasing check with if constexpr blocks
- Created small-cache optimized instantiations (MaxNumBlocks=100000)
- Added runtime dispatch logic for small vs large cache
- 3.7% performance improvement for small caches vs runtime check

2026-05-07 07:43:48 +00:00

..

ck

[rocm-libraries] ROCm/rocm-libraries#6022 (commit 54b284a)

2026-03-31 15:19:43 +00:00

ck_tile

Add compile-time MaxNumBlocks optimization

2026-05-07 07:43:48 +00:00

rapidjson

…