diff --git a/src/ggml-dnnl.cpp b/src/ggml-dnnl.cpp
index b43b452693..031b3eae8a 100644
--- a/src/ggml-dnnl.cpp
+++ b/src/ggml-dnnl.cpp
@@ -877,7 +877,7 @@ static void ggml_backend_dnnl_buffer_init_tensor(ggml_backend_buffer_t buffer, g
     if (tensor->view_src != NULL && tensor->view_offs == 0) {
         tensor->extra = tensor->view_src->extra;
     } else {
-    printf(" op:%s-'%s'\n", ggml_op_desc(tensor), tensor->name);
+    //printf(" op:%s-'%s'\n", ggml_op_desc(tensor), tensor->name);
         auto buf = tensor->view_src != NULL ? tensor->view_src->buffer : tensor->buffer;
         ggml_backend_dnnl_buffer_context* ctx = (ggml_backend_dnnl_buffer_context*)buf->context;
         dnnl::memory::dim offset = (uintptr_t)tensor->data - DNNL_BUFFER_BASE;
@@ -976,7 +976,6 @@ GGML_CALL static ggml_backend_buffer_type_t ggml_backend_dnnl_get_default_buffer
             /* .get_alignment    = */ ggml_backend_dnnl_buffer_type_get_alignment,
             /* .get_max_size     = */ NULL, // defaults to SIZE_MAX
             /* .get_alloc_size   = */ NULL, // defaults to ggml_nbytes
-            /* .supports_backend = */ ggml_backend_dnnl_buffer_type_supports_backend,
             /* .is_host          = */ NULL, // defaults to false // ggml_backend_dnnl_buffer_type_is_host,
         },
         /* .context = */ backend->context,
@@ -993,12 +992,6 @@ GGML_CALL static const char * ggml_backend_dnnl_buffer_type_get_name(ggml_backen
     GGML_UNUSED(buft);
 }
 
-GGML_CALL static bool ggml_backend_dnnl_buffer_type_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend) {
-    return ggml_backend_is_dnnl(backend) || ggml_backend_is_cpu(backend);
-
-    GGML_UNUSED(buft);
-}
-
 GGML_CALL static bool ggml_backend_dnnl_buffer_type_is_host(ggml_backend_buffer_type_t buft) {
     return true;
 
@@ -1015,7 +1008,6 @@ GGML_CALL static ggml_backend_buffer_type_t ggml_backend_dnnl_get_default_buffer
             /* .get_alignment    = */ cpu_buffer_type->iface.get_alignment, // ggml_backend_cpu_buffer_type_get_alignment,
             /* .get_max_size     = */ NULL, // defaults to SIZE_MAX
             /* .get_alloc_size   = */ NULL, // defaults to ggml_nbytes
-            /* .supports_backend = */ ggml_backend_dnnl_buffer_type_supports_backend,
             /* .is_host          = */ ggml_backend_dnnl_buffer_type_is_host,
         },
         /* .context = */ NULL,
@@ -1067,6 +1059,20 @@ GGML_CALL static bool ggml_backend_dnnl_supports_op(ggml_backend_t backend, cons
     return ggml_backend_dnnl_node_supported(backend, op);
 }
 
+static bool ggml_backend_buft_is_dnnl(ggml_backend_buffer_type_t buft) {
+    return buft->iface.get_name == ggml_backend_dnnl_buffer_type_get_name;
+}
+
+GGML_CALL static bool ggml_backend_dnnl_supports_buft(ggml_backend_t backend, ggml_backend_buffer_type_t buft) {
+    GGML_UNUSED(backend);
+#if USE_DNNL_BUFFER
+    return ggml_backend_buft_is_dnnl(buft);
+#else
+    return ggml_backend_buft_is_dnnl(buft) || ggml_backend_buft_is_host(buft);
+#endif
+}
+
+
 static struct ggml_backend_i dnnl_backend_i = {
     /* .get_name                = */ ggml_backend_dnnl_name,
     /* .free                    = */ ggml_backend_dnnl_free,
@@ -1077,15 +1083,17 @@ static struct ggml_backend_i dnnl_backend_i = {
     /* .synchronize             = */ ggml_backend_dnnl_synchronize,
     /* .graph_plan_create       = */ NULL,
     /* .graph_plan_free         = */ NULL,
+    /* .graph_plan_update       = */ NULL,
     /* .graph_plan_compute      = */ NULL,
     /* .graph_compute           = */ ggml_backend_dnnl_graph_compute,
     /* .supports_op             = */ ggml_backend_dnnl_supports_op,
+    /* .supports_buft           = */ ggml_backend_dnnl_supports_buft,
     /* .offload_op              = */ NULL,
     /* .event_new               = */ NULL,
     /* .event_free              = */ NULL,
     /* .event_record            = */ NULL,
     /* .event_wait              = */ NULL,
-    /* .event_synchronize       = */ NULL,
+    /* .event_synchronize       = */ NULL
 };
 
 static ggml_guid_t ggml_backend_dnnl_guid() {