From e64f237e04a3fc5fbf14dab802325871769f7cdc Mon Sep 17 00:00:00 2001
From: Lizonghang <870644199@qq.com>
Date: Sat, 15 Feb 2025 17:43:03 +0400
Subject: [PATCH] fix bugs in available_mem calculation

---
 common/common.cpp | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/common/common.cpp b/common/common.cpp
index bcd3c49a..f13a2e12 100644
--- a/common/common.cpp
+++ b/common/common.cpp
@@ -1432,7 +1432,11 @@ static bool assign_layers_to_device(
 
         if (dev.gpu_support.cuda || dev.gpu_support.metal) {
             int64_t required_mem = w[m] * b_prime;
-            int64_t available_mem = dev.gpu_props.memory_free * GIGABYTE - c_gpu[m];            
+            int64_t available_mem = dev.gpu_props.memory_free * GIGABYTE - c_gpu[m];
+            if (dev.gpu_support.metal && m == 0 && cparams.keep_out_in_metal) {
+                available_mem -= bo;
+            }
+
             if (required_mem <= available_mem) {
                 n_gpu_layers[m] = w[m];
             } else {