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

post review

Browse source
This commit is contained in:
Y. Velkov 2024-05-16 09:57:32 +03:00
parent 89b3236152
commit e02d3d0c29
2 changed files with 7 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

15
ggml-cuda.cu
View file

@ -51,14 +51,9 @@
static_assert(sizeof(half) == sizeof(ggml_fp16_t), "wrong fp16 size");
static void ggml_cuda_default_log_callback(enum ggml_log_level level, const char * msg, void * user_data) {
GGML_UNUSED(level);
GGML_UNUSED(user_data);
if (level == GGML_LOG_LEVEL_WARN) {
fprintf(stderr, "warning: %s", msg);
} else if (level == GGML_LOG_LEVEL_ERROR) {
fprintf(stderr, "error: %s", msg);
} else {
fprintf(stderr, "%s", msg);
}
fprintf(stderr, "%s", msg);
}
ggml_log_callback ggml_cuda_log_callback = ggml_cuda_default_log_callback;
@ -98,9 +93,9 @@ void ggml_cuda_error(const char * stmt, const char * func, const char * file, in
int id = -1; // in case cudaGetDevice fails
cudaGetDevice(&id);
GGML_CUDA_LOG_INFO("CUDA error: %s\n", msg);
GGML_CUDA_LOG_INFO(" current device: %d, in function %s at %s:%d\n", id, func, file, line);
GGML_CUDA_LOG_INFO(" %s\n", stmt);
GGML_CUDA_LOG_ERROR("CUDA error: %s\n", msg);
GGML_CUDA_LOG_ERROR(" current device: %d, in function %s at %s:%d\n", id, func, file, line);
GGML_CUDA_LOG_ERROR(" %s\n", stmt);
// abort with GGML_ASSERT to get a stack trace
GGML_ASSERT(!"CUDA error");
}

2
llama.cpp
View file

@ -1778,6 +1778,8 @@ struct llama_state {
llama_state() {
#ifdef GGML_USE_METAL
ggml_backend_metal_log_set_callback(log_callback, log_callback_user_data);
#elif GGML_USE_CUDA
ggml_backend_cuda_log_set_callback(log_callback, log_callback_user_data);
#endif
}