From a5a32b9179a27828aff3c72d2fe4c3f8dc8e69e2 Mon Sep 17 00:00:00 2001
From: Nexesenex <124105151+Nexesenex@users.noreply.github.com>
Date: Sat, 29 Jun 2024 04:10:51 +0200
Subject: [PATCH] Add Cuda Graphs in CMakeList (#947)

original LCPP PR 3242 by Slaren
---
 CMakeLists.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index a83ec68b0..d086f2065 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -49,6 +49,7 @@ option(LLAMA_CUDA_F16                        "llama: use 16 bit floats for dmmv
 set(LLAMA_CUDA_KQUANTS_ITER "2" CACHE STRING "llama: iters./thread per block for Q2_K/Q6_K")
 set(LLAMA_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING
                                              "llama: max. batch size for using peer access")
+set(GGML_CUDA_USE_GRAPHS                                                                        ON)
 option(LLAMA_HIPBLAS                         "llama: use hipBLAS"                               OFF)
 
 # Other