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

blasthreads should apply for any thread count below 32

Browse source
This commit is contained in:
Concedo 2023-10-08 15:51:18 +08:00
parent a2b8473354
commit 80dfe2ba49
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama.cpp
View file

@ -4972,7 +4972,7 @@ static int llama_decode_internal(
GGML_ASSERT(n_tokens <= n_batch);
int n_threads = n_tokens == 1 ? cparams.n_threads : cparams.n_threads_batch;
int n_threads = n_tokens < 32 ? cparams.n_threads : cparams.n_threads_batch;
GGML_ASSERT((!batch.token && batch.embd) || (batch.token && !batch.embd)); // NOLINT
const int64_t t_start_us = ggml_time_us();