Back to feed

b7549

Dec 26, 2025
Meta/llama.cppCLIvb7549

vulkan: preprocess mul_mat_id experts and discard workgroups more quickly (#18352)

Run a preprocess to count how many times each expert is used, and use this to quickly discard workgroups that aren't needed.

macOS/iOS:

Linux:

Windows:

openEuler: