llama.cpp
Vulkan: Optimize Matmul parameters for AMD GPUs with Coopmat support
#18749
Merged
Overview
Commits
3
Changes
View On GitHub
Commits
  • vulkan: Enable and optimize large matmul parameter combination for AMD
    0cc4m committed 6 days ago
  • limit tuning to AMD GPUs with coopmat support
    0cc4m committed 5 days ago
  • use tx_m values instead of _l
    0cc4m committed 5 days ago
FAQ Terms Privacy Refunds Impressum
Loading