llama.cpp

58546250 - opencl: add bin kernels `kernel_gemm_moe_q4_0_q8_1_dp4a_bin`, `kernel_gemm_moe_mxfp4_q8_1_dp4a_bin` (#27768)