From 20ca856ab1cc73582c556b69e7572a7ea64718e5 Mon Sep 17 00:00:00 2001
From: slaren <slarengh@gmail.com>
Date: Mon, 7 Oct 2024 22:45:30 +0200
Subject: [PATCH] llama : print devices used on model load

---
 ggml/src/ggml-backend.cpp | 3 +--
 src/llama.cpp             | 5 +++++
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/ggml/src/ggml-backend.cpp b/ggml/src/ggml-backend.cpp
index 391af6a92..16ccd2595 100644
--- a/ggml/src/ggml-backend.cpp
+++ b/ggml/src/ggml-backend.cpp
@@ -555,7 +555,6 @@ struct ggml_backend_registry {
         register_backend(ggml_backend_cuda_reg());
 #endif
 #ifdef GGML_USE_METAL
-
         register_backend(ggml_backend_metal_reg());
 #endif
 #ifdef GGML_USE_VULKAN
@@ -565,7 +564,7 @@ struct ggml_backend_registry {
         register_backend(ggml_backend_blas_reg());
 #endif
 
-        // TODO: sycl, vulkan, kompute, cann
+        // TODO: sycl, kompute, cann
 
         register_backend(ggml_backend_cpu_reg());
     }
diff --git a/src/llama.cpp b/src/llama.cpp
index 3a49ecffb..9e83c7913 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -19100,8 +19100,13 @@ struct llama_model * llama_load_model_from_file(
 
             case GGML_BACKEND_DEVICE_TYPE_GPU:
             case GGML_BACKEND_DEVICE_TYPE_GPU_FULL:
+            {
+                size_t free, total; // NOLINT
+                ggml_backend_dev_memory(dev, &free, &total);
+                LLAMA_LOG_INFO("%s: using device %s (%s) - %zu MiB free\n", __func__, ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), free/1024/1024);
                 model->devices.push_back(dev);
                 break;
+            }
         }
     }