llama.cpp
c9ced491
- vulkan: preprocess mul_mat_id experts and discard workgroups more quickly (#18352)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
28 days ago
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.
References
#18352 - vulkan: preprocess mul_mat_id experts and discard workgroups more quickly
Author
jeffbolznv
Parents
7ac89021
Loading