From: Johannes Gäßler <redacted>
Date: Wed, 24 Sep 2025 14:53:48 +0000 (+0200)
Subject: llama: print memory breakdown on exit (llama/15860)
X-Git-Tag: v0.9.3~10
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=1a2126fd856b7de0f364cffa1730851e8a7ebadd;p=pkg%2Fggml%2Fsources%2Fggml

llama: print memory breakdown on exit (llama/15860)

* llama: print memory breakdown on exit
---

diff --git a/include/ggml-backend.h b/include/ggml-backend.h
index ab297e0c..62b6d65e 100644
--- a/include/ggml-backend.h
+++ b/include/ggml-backend.h
@@ -314,7 +314,8 @@ extern "C" {
     GGML_API int                  ggml_backend_sched_get_n_splits(ggml_backend_sched_t sched);
     GGML_API int                  ggml_backend_sched_get_n_copies(ggml_backend_sched_t sched);
 
-    GGML_API size_t               ggml_backend_sched_get_buffer_size(ggml_backend_sched_t sched, ggml_backend_t backend);
+    GGML_API ggml_backend_buffer_type_t ggml_backend_sched_get_buffer_type(ggml_backend_sched_t sched, ggml_backend_t backend);
+    GGML_API size_t                     ggml_backend_sched_get_buffer_size(ggml_backend_sched_t sched, ggml_backend_t backend);
 
     GGML_API void                 ggml_backend_sched_set_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node, ggml_backend_t backend);
     GGML_API ggml_backend_t       ggml_backend_sched_get_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node);
diff --git a/src/ggml-backend.cpp b/src/ggml-backend.cpp
index 79a5282b..ff9135fe 100644
--- a/src/ggml-backend.cpp
+++ b/src/ggml-backend.cpp
@@ -1793,6 +1793,14 @@ ggml_backend_t ggml_backend_sched_get_backend(ggml_backend_sched_t sched, int i)
     return sched->backends[i];
 }
 
+ggml_backend_buffer_type_t ggml_backend_sched_get_buffer_type(ggml_backend_sched_t sched, ggml_backend_t backend) {
+    GGML_ASSERT(sched);
+    int backend_index = ggml_backend_sched_backend_id(sched, backend);
+    GGML_ASSERT(backend_index >= 0 && backend_index < sched->n_backends);
+
+    return sched->bufts[backend_index];
+}
+
 size_t ggml_backend_sched_get_buffer_size(ggml_backend_sched_t sched, ggml_backend_t backend) {
     GGML_ASSERT(sched);
     int backend_index = ggml_backend_sched_backend_id(sched, backend);