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

llama : avoid unnecessary bool

Browse source
This commit is contained in:
Georgi Gerganov 2023-07-06 21:54:04 +03:00
parent 551ed08234
commit f789f2cef2
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
1 changed files with 3 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

9
llama.cpp
View file

@ -1613,14 +1613,11 @@ static bool llama_eval_internal(
// run the computation
ggml_build_forward_expand(&gf, cur);
bool call_ggml_graph_compute = true;
#ifdef GGML_USE_METAL
if (lctx.ctx_metal && N == 1) {
ggml_metal_set_n_cb (lctx.ctx_metal, n_threads);
ggml_metal_graph_compute(lctx.ctx_metal, &gf);
ggml_metal_get_tensor (lctx.ctx_metal, cur);
call_ggml_graph_compute = false;
} else {
// IMPORTANT:
// Since we don't have efficient Matrix x Matrix Metal multiplication yet, we fallback to vanilla
@ -1637,12 +1634,12 @@ static bool llama_eval_internal(
ggml_metal_get_tensor(lctx.ctx_metal, kv_self.k);
ggml_metal_get_tensor(lctx.ctx_metal, kv_self.v);
}
}
#endif
if (call_ggml_graph_compute) {
ggml_graph_compute_helper(lctx.work_buffer, &gf, n_threads);
}
#else
ggml_graph_compute_helper(lctx.work_buffer, &gf, n_threads);
#endif
if (cgraph_fname) {
ggml_graph_export(&gf, cgraph_fname);