From 2dd464760de752476e76bbe145afa2fd109b07f6 Mon Sep 17 00:00:00 2001
From: leejet <leejet714@gmail.com>
Date: Sat, 12 Sep 2026 01:02:28 +0800
Subject: [PATCH] fix: reuse graph plans when scale parameters change

---
 src/core/ggml_graph_cut.cpp | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/src/core/ggml_graph_cut.cpp b/src/core/ggml_graph_cut.cpp
index 797100587..a6e2d99e9 100644
--- a/src/core/ggml_graph_cut.cpp
+++ b/src/core/ggml_graph_cut.cpp
@@ -482,6 +482,16 @@ namespace sd::ggml_graph_cut {
         return ggml_nbytes(cache_src);
     }
 
+    static bool can_ignore_op_params(ggml_op op) {
+        // Exempt only parameters that cannot affect graph layout or backend allocation size.
+        switch (op) {
+            case GGML_OP_SCALE:
+                return true;
+            default:
+                return false;
+        }
+    }
+
     std::vector<uint64_t> graph_layout(ggml_cgraph* graph, bool include_bindings) {
         std::vector<const ggml_tensor*> tensors;
         std::unordered_map<const ggml_tensor*, size_t> indices;
@@ -530,8 +540,10 @@ namespace sd::ggml_graph_cut {
             for (auto source : tensor->src) {
                 signature.push_back(source == nullptr ? 0 : indices.at(source));
             }
-            for (int value : tensor->op_params) {
-                signature.push_back(static_cast<uint32_t>(value));
+            if (!can_ignore_op_params(tensor->op)) {
+                for (int value : tensor->op_params) {
+                    signature.push_back(static_cast<uint32_t>(value));
+                }
             }
         }
         return signature;