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

fix build break on MacOS, due to CI of MacOS depend on external ggml, instead of internal ggml

Browse source
This commit is contained in:
jianyuzh 2024-01-24 17:24:24 +08:00
parent ec5c8bc0c9
commit 22e1b45c02
1 changed files with 5 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

6
llama.cpp
View file

@ -6703,10 +6703,14 @@ static int llama_decode_internal(
} }
const bool fully_offloaded = model.n_gpu_layers >= (int) hparams.n_layer + 1; const bool fully_offloaded = model.n_gpu_layers >= (int) hparams.n_layer + 1;
if ((ggml_cpu_has_cublas() || ggml_cpu_has_sycl()) && fully_offloaded) { if (ggml_cpu_has_cublas() && fully_offloaded) {
n_threads = 1; n_threads = 1;
} }
#ifdef GGML_USE_SYCL
n_threads = 1;
#endif
#ifdef GGML_USE_MPI #ifdef GGML_USE_MPI
const int64_t n_layer = hparams.n_layer; const int64_t n_layer = hparams.n_layer;
ggml_mpi_graph_compute_pre(lctx.ctx_mpi, gf, n_layer); ggml_mpi_graph_compute_pre(lctx.ctx_mpi, gf, n_layer);