Loosen test_point_face_distance grad_faces tolerance to 5e-6

DistanceBackwardCuda accumulates per-face gradient contributions via atomicAdd and explicitly calls `at::globalContext().alertNotDeterministic("DistanceBackwardCuda")`. The FP add order is platform-dependent — different wavefront/warp widths and scheduling produce different rounding errors. On AMD gfx90a (warpSize=64, MI250X) the resulting grad_faces value differs from the CPU autograd naive reference by a bit-stable 7.67e-7 max abs (4e-5 max rel), which exceeds the previous `atol=5e-7` but is well within float32 noise.

The sibling test `test_face_point_distance` (5 lines below) already uses `atol=5e-6` for exactly the same kind of comparison on grad_tris; the asymmetric `5e-7` for this test was an inconsistency rather than a hardware-specific requirement. Bumping both grad_faces asserts (cuda and cpu paths) to `5e-6` brings the two tests into alignment and accepts platforms with different atomic-ordering behaviour.

Co-Authored-By: Claude Opus 4.7 (1M context)

Author: jeffdaily

tests/test_point_mesh_distance.py

@@ -707,9 +707,14 @@ def test_point_face_distance(self):
 
         # Compare
         self.assertClose(grad_points_naive.cpu(), grad_points_cuda.cpu(), atol=1e-7)
-        self.assertClose(grad_faces_naive, grad_faces_cuda.cpu(), atol=5e-7)
+        # DistanceBackward uses atomicAdd to accumulate gradients into the
+        # face buffer and explicitly calls alertNotDeterministic; the FP add
+        # order differs across GPU architectures (e.g. between 32- and
+        # 64-lane warps), producing tiny rounding differences. Use the same
+        # 5e-6 tolerance as test_face_point_distance below.
+        self.assertClose(grad_faces_naive, grad_faces_cuda.cpu(), atol=5e-6)
         self.assertClose(grad_points_naive.cpu(), grad_points_cpu, atol=1e-7)
-        self.assertClose(grad_faces_naive, grad_faces_cpu, atol=5e-7)
+        self.assertClose(grad_faces_naive, grad_faces_cpu, atol=5e-6)
 
     def test_face_point_distance(self):
         """