Merge branch 'ershi/more-release-1.3.2' into 'release-1.3'

More Release 1.3.2 Fixes

See merge request omniverse/warp!708

CHANGELOG.md

@@ -5,7 +5,7 @@
 - Bug fixes
  - Fix accuracy of 3x3 SVD ``wp.svd3`` with fp64 numbers ([GH-281](https://github.com/NVIDIA/warp/issues/281)).
  - Fix module hashing when a kernel argument contained a struct array ([GH-287](https://github.com/NVIDIA/warp/issues/287)).
- - Fix a bug in `wp.bvh_query_aabb()` where the direction instead of the reciprocal direction was used
+ - Fix a bug in `wp.bvh_query_ray()` where the direction instead of the reciprocal direction was used
  ([GH-288](https://github.com/NVIDIA/warp/issues/288)).
  - Fix errors when launching a CUDA graph after a module is reloaded. Modules that were used during graph capture
  will no longer be unloaded before the graph is released.
@@ -18,10 +18,11 @@
  - Fix `ModelBuilder.collapse_fixed_joints()` to correctly update the body centers of mass and the
  `ModelBuilder.articulation_start` array.
  - Fix precedence of closure constants over global constants.
+ - Fix quadrature point indexing in `wp.fem.ExplicitQuadrature` (regression from 1.3.0).
 - Documentation improvements
  - Add missing return types for built-in functions.
  - Clarify that atomic operations also return the previous value.
- - Clarify that `wp.bvh_query_aabb` returns parts that overlap the bounding volume.
+ - Clarify that `wp.bvh_query_aabb()` returns parts that overlap the bounding volume.
 
 ## [1.3.1] - 2024-07-27
 

exts/omni.warp.core/docs/CHANGELOG.md

@@ -5,7 +5,7 @@
 - Bug fixes
  - Fix accuracy of 3x3 SVD ``wp.svd3`` with fp64 numbers ([GH-281](https://github.com/NVIDIA/warp/issues/281)).
  - Fix module hashing when a kernel argument contained a struct array ([GH-287](https://github.com/NVIDIA/warp/issues/287)).
- - Fix a bug in `wp.bvh_query_aabb()` where the direction instead of the reciprocal direction was used
+ - Fix a bug in `wp.bvh_query_ray()` where the direction instead of the reciprocal direction was used
  ([GH-288](https://github.com/NVIDIA/warp/issues/288)).
  - Fix errors when launching a CUDA graph after a module is reloaded. Modules that were used during graph capture
  will no longer be unloaded before the graph is released.
@@ -18,10 +18,11 @@
  - Fix `ModelBuilder.collapse_fixed_joints()` to correctly update the body centers of mass and the
  `ModelBuilder.articulation_start` array.
  - Fix precedence of closure constants over global constants.
+ - Fix quadrature point indexing in `wp.fem.ExplicitQuadrature` (regression from 1.3.0).
 - Documentation improvements
  - Add missing return types for built-in functions.
  - Clarify that atomic operations also return the previous value.
- - Clarify that `wp.bvh_query_aabb` returns parts that overlap the bounding volume.
+ - Clarify that `wp.bvh_query_aabb()` returns parts that overlap the bounding volume.
 
 ## [1.3.1] - 2024-07-27
 

exts/omni.warp/docs/CHANGELOG.md

@@ -5,7 +5,7 @@
 - Bug fixes
  - Fix accuracy of 3x3 SVD ``wp.svd3`` with fp64 numbers ([GH-281](https://github.com/NVIDIA/warp/issues/281)).
  - Fix module hashing when a kernel argument contained a struct array ([GH-287](https://github.com/NVIDIA/warp/issues/287)).
- - Fix a bug in `wp.bvh_query_aabb()` where the direction instead of the reciprocal direction was used
+ - Fix a bug in `wp.bvh_query_ray()` where the direction instead of the reciprocal direction was used
  ([GH-288](https://github.com/NVIDIA/warp/issues/288)).
  - Fix errors when launching a CUDA graph after a module is reloaded. Modules that were used during graph capture
  will no longer be unloaded before the graph is released.
@@ -18,10 +18,11 @@
  - Fix `ModelBuilder.collapse_fixed_joints()` to correctly update the body centers of mass and the
  `ModelBuilder.articulation_start` array.
  - Fix precedence of closure constants over global constants.
+ - Fix quadrature point indexing in `wp.fem.ExplicitQuadrature` (regression from 1.3.0).
 - Documentation improvements
  - Add missing return types for built-in functions.
  - Clarify that atomic operations also return the previous value.
- - Clarify that `wp.bvh_query_aabb` returns parts that overlap the bounding volume.
+ - Clarify that `wp.bvh_query_aabb()` returns parts that overlap the bounding volume.
 
 ## [1.3.1] - 2024-07-27
 

warp/fem/quadrature/quadrature.py

@@ -336,6 +336,7 @@ def max_points_per_element(self):
  @cache.cached_arg_value
  def arg_value(self, device):
  arg = self.Arg()
+ arg.points_per_cell = self._points_per_cell
  arg.points = self._points.to(device)
  arg.weights = self._weights.to(device)
 

warp/tests/test_fem.py

@@ -1252,6 +1252,8 @@ def test_particle_quadratures(test, device):
  geo = fem.Grid2D(res=wp.vec2i(2))
 
  domain = fem.Cells(geo)
+
+ # Explicit quadrature
  points, weights = domain.reference_element().instantiate_quadrature(order=4, family=fem.Polynomial.GAUSS_LEGENDRE)
  points_per_cell = len(points)
 
@@ -1266,9 +1268,16 @@ def test_particle_quadratures(test, device):
  test.assertEqual(explicit_quadrature.max_points_per_element(), points_per_cell)
  test.assertEqual(explicit_quadrature.total_point_count(), points_per_cell * geo.cell_count())
 
+ # test integration accuracy
  val = fem.integrate(_bicubic, quadrature=explicit_quadrature)
  test.assertAlmostEqual(val, 1.0 / 16, places=5)
 
+ # test indexing validity
+ arr = wp.empty(explicit_quadrature.total_point_count(), dtype=float)
+ fem.interpolate(_piecewise_constant, dest=arr, quadrature=explicit_quadrature)
+ assert_np_equal(arr.numpy(), np.arange(geo.cell_count()).repeat(points_per_cell))
+
+ # PIC quadrature
  element_indices = wp.array([3, 3, 2], dtype=int, device=device)
  element_coords = wp.array(
  [
@@ -1286,6 +1295,7 @@ def test_particle_quadratures(test, device):
  test.assertEqual(pic_quadrature.total_point_count(), 3)
  test.assertEqual(pic_quadrature.active_cell_count(), 2)
 
+ # Test integration accuracy
  val = fem.integrate(_piecewise_constant, quadrature=pic_quadrature)
  test.assertAlmostEqual(val, 1.25, places=5)