root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-25 10:54:36 +00:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

llama : disable pipeline parallelism with nkvo (#7265)

Browse Source
This commit is contained in:
slaren 2024-05-14 09:33:42 +02:00 committed by GitHub
parent efc8f767c8
commit 541600201e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
llama.cpp
View File

@ -15849,7 +15849,11 @@ struct llama_context * llama_new_context_with_model(
ctx->buf_compute_meta.resize(ggml_tensor_overhead()*LLAMA_MAX_NODES + ggml_graph_overhead_custom(LLAMA_MAX_NODES, false)); ctx->buf_compute_meta.resize(ggml_tensor_overhead()*LLAMA_MAX_NODES + ggml_graph_overhead_custom(LLAMA_MAX_NODES, false));
// enabling pipeline parallelism in the scheduler increases memory usage, so it is only done when necessary // enabling pipeline parallelism in the scheduler increases memory usage, so it is only done when necessary
bool pipeline_parallel = llama_get_device_count() > 1 && model->n_gpu_layers > (int)model->hparams.n_layer && model->split_mode == LLAMA_SPLIT_MODE_LAYER; bool pipeline_parallel =
llama_get_device_count() > 1 &&
model->n_gpu_layers > (int)model->hparams.n_layer &&
model->split_mode == LLAMA_SPLIT_MODE_LAYER &&
params.offload_kqv;
#ifndef GGML_USE_CUDA #ifndef GGML_USE_CUDA
// pipeline parallelism requires support for async compute and events // pipeline parallelism requires support for async compute and events
// currently this is only implemented in the CUDA backend // currently this is only implemented in the CUDA backend