vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-12 01:54:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Merge branch 'master' into concedo_experimental

Browse source 
# Conflicts:
#	CMakeLists.txt
#	Makefile
#	README.md
#	ggml.c
This commit is contained in:
Concedo 2023-04-29 11:14:05 +08:00
commit 0fc1772a8f
25 changed files with 1156 additions and 740 deletions
Download patch file Download diff file Expand all files Collapse all files

26
llama_util.h
View file

@ -405,4 +405,30 @@ struct llama_buffer {
delete[] addr;
}
};
#ifdef GGML_USE_CUBLAS
#include "ggml-cuda.h"
struct llama_ctx_buffer {
uint8_t * addr = NULL;
size_t size = 0;
void resize(size_t size) {
if (addr) {
ggml_cuda_host_free(addr);
}
addr = (uint8_t *) ggml_cuda_host_malloc(size);
this->size = size;
}
~llama_ctx_buffer() {
if (addr) {
ggml_cuda_host_free(addr);
}
}
};
#else
typedef llama_buffer llama_ctx_buffer;
#endif
#endif