llama.cpp
batched-bench : fix llama_synchronize usage during prompt processing
#15835
Merged
Merged