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

Multi GPU support, CUDA refactor, CUDA scratch buffer (#1703)

Browse source 
* CUDA multi GPU + scratch

ggml_cuda_compute_forward

Tensor parallelism

ggml_cuda_add

ggml_cuda_rms_norm

ggml_cuda_silu

CUDA scratch buffer

--main-gpu CLI option
This commit is contained in:
Johannes Gäßler 2023-06-06 21:33:23 +02:00 • committed by GitHub
parent 44f906e853
commit 17366df842
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
12 changed files with 1221 additions and 544 deletions
Download patch file Download diff file Expand all files Collapse all files

16
llama.h
View file

@ -1,6 +1,13 @@
#ifndef LLAMA_H
#define LLAMA_H
#include "ggml.h"
#ifdef GGML_USE_CUBLAS
#include "ggml-cuda.h"
#define LLAMA_MAX_DEVICES GGML_CUDA_MAX_DEVICES
#else
#define LLAMA_MAX_DEVICES 1
#endif // GGML_USE_CUBLAS
#include <stddef.h>
#include <stdint.h>
#include <stdbool.h>
@ -65,9 +72,12 @@ extern "C" {
typedef void (*llama_progress_callback)(float progress, void *ctx);
struct llama_context_params {
int n_ctx; // text context
int n_gpu_layers; // number of layers to store in VRAM
int seed; // RNG seed, -1 for random
int n_ctx; // text context
int n_batch; // prompt processing batch size
int n_gpu_layers; // number of layers to store in VRAM
int main_gpu; // the GPU that is used for scratch and small tensors
float tensor_split[LLAMA_MAX_DEVICES]; // how to split layers across multiple GPUs
int seed; // RNG seed, -1 for random
bool f16_kv; // use fp16 for KV cache
bool logits_all; // the llama_eval() call computes all logits, not just the last one