b10357 b10356

ggml-org/llama.cppAug 11, 2026 → Aug 11, 20260 days between releases

Generating AI comparison

Analyzing what changed between b10357 and b10356. This page updates automatically — usually within a minute.

The Two Releases

b10357

b10357

Aug 11, 2026· 2 months ago

Optimized OpenCL Flash Attention prefill kernels by transposing the K tile in local memory to improve performance.

Full release notes
b10356

b10356

Aug 11, 2026· 2 months ago

Updated CI infrastructure to target ROCm 7.14, adopting the new TheRock build system and migrating Windows ROCm jobs to use multi-arch wheels.

Full release notes