vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

revert no mmq

Browse source
This commit is contained in:
Meng, Hengyu 2024-01-24 12:37:29 +00:00
parent 8dd1b60ac6
commit 3aabd8a274
1 changed files with 1 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

9
common/common.cpp
View file

@ -631,7 +631,6 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
invalid_param = true;
break;
}
#if defined(GGML_USE_CUBLAS) || defined(GGML_USE_SYCL)
std::string arg_next = argv[i];
// split string by , and /
@ -649,15 +648,9 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
params.tensor_split[i] = 0.0f;
}
}
#else
#ifndef GGML_USE_CUBLAS_SYCL
fprintf(stderr, "warning: llama.cpp was compiled without cuBLAS/SYCL. Setting a tensor split has no effect.\n");
#endif // GGML_USE_CUBLAS_SYCL
} else if (arg == "--no-mul-mat-q" || arg == "-nommq") {
#if defined(GGML_USE_CUBLAS) || defined(GGML_USE_SYCL)
params.mul_mat_q = false;
#else
fprintf(stderr, "warning: llama.cpp was compiled without cuBLAS or SYCL. Disabling mul_mat_q kernels has no effect.\n");
#endif // GGML_USE_CUBLAS || GGML_USE_SYCL
} else if (arg == "--no-mmap") {
params.use_mmap = false;
} else if (arg == "--numa") {