llama.cpp

cuda : update supports_op for matrix multiplication

#8245

Merged