From 5f022185a1170b2cf1eec37dd07b6db8638a5d88 Mon Sep 17 00:00:00 2001
From: netrunnereve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 5 Aug 2023 22:39:44 -0400
Subject: [PATCH] test pp_threads

---
 llama.cpp | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/llama.cpp b/llama.cpp
index 839739870..a44f648df 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -1814,9 +1814,12 @@ static bool llama_eval_internal(
 
     // fprintf(stderr, "graph build time: %.3f ms (%d nodes, %d leafs)\n", (ggml_time_us() - t_start_us)/1000.0, gf->n_nodes, gf->n_leafs);
 
+    int32_t pp_threads = 3;
+
     // for big prompts, if BLAS is enabled, it is better to use only one thread
     // otherwise, the threads are spin-lock waiting for the BLAS calls and are degrading the performance
-    n_threads = N >= 32 && ggml_cpu_has_blas() && !ggml_cpu_has_gpublas() ? 1 : n_threads;
+    pp_threads = N >= 32 && ggml_cpu_has_blas() && !ggml_cpu_has_gpublas() ? 1 : pp_threads;
+    n_threads = N > 1 ? pp_threads : n_threads;
 
     struct ggml_tensor * res = gf->nodes[gf->n_nodes - 1];
     struct ggml_tensor * embeddings = gf->nodes[gf->n_nodes - 2];