llama.cpp
vulkan: allow graph_optimize for prompt processing workloads
#17475
Merged
Merged