refactor: vectorize gradient loop and fix DD node position

Replace per-pixel Python loop with vectorized torch.arange + slice
operations. Fix DifferentialDiffusion node position to avoid visual
overlap with SplitImageToTileList node 14 on the canvas.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

example_workflows/tiled-upscale-builtin-nodes.json

@@ -412,7 +412,7 @@
     {
       "id": 24,
       "type": "DifferentialDiffusion",
-      "pos": [2560, 160],
+      "pos": [2560, 60],
       "size": [250, 46],
       "flags": {},
       "order": 12,