Issues

• fvdb.nn.SparseConv3d.weight is a non-contiguous view; document it and give flatten-based trainers a fast path

  documentationoptimizationPython APItriage
  openvdb/fvdb-core#832 · Filter by author swahtz opened on Oct 21, 2026

• refined_grid should accept a per-child subdivision mask

  enhancementPython APITopology Operationstriage
  openvdb/fvdb-core#831 · Filter by author swahtz opened on Oct 21, 2026

• Extend VoxelBlockManager iteration across CUDA ops (follow-ups to #747)

  core libraryenhancementoptimizationtriage
  openvdb/fvdb-core#817 · Filter by author swahtz opened on Oct 13, 2026

• CUDA jcat(dim=0) writes wrong jidx and reads one past the end of offsets

  bugcore librarytriage
  openvdb/fvdb-core#811 · Filter by author swahtz opened on Oct 8, 2026
  

• JaggedTensor.lshape segfaults on ldim-2 tensor built from empty list_ids with num_tensors > 0

  bugcore librarytriage
  openvdb/fvdb-core#805 · Filter by author harrism opened on Oct 6, 2026
  

• rasterize_contributing_gaussian_ids_sparse segfaults when no requested pixel has a contributor

  bugGaussian Splattingtriage
  openvdb/fvdb-core#803 · Filter by author swahtz opened on Sep 23, 2026
  

• JaggedTensor: empty outer lists are dropped by from_data_offsets_and_list_ids, hidden from lshape/unbind, and fail CPU indexing

  bugcore libraryPython APItriage
  openvdb/fvdb-core#802 · Filter by author swahtz opened on Sep 23, 2026
  1

  

• Dense rasterizers cannot render a crop into crop-sized buffers: window origin and tile-grid contract are inconsistent

  bugGaussian Splattingtriage
  openvdb/fvdb-core#800 · Filter by author swahtz opened on Sep 22, 2026

• voxel_size and origin are rounded to float32 on the way into every grid constructor

  bugPython APItriage
  openvdb/fvdb-core#798 · Filter by author swahtz opened on Sep 22, 2026

• Delete the deprecated TorchDeviceBuffer shim at the NanoVDB pin bump past the dual-space removal

  triage
  openvdb/fvdb-core#792 · Filter by author harrism opened on Sep 17, 2026

• Batch NanoVDB's TopologyBuilder upstream and make it fvdb's batched topology engine

  triage
  openvdb/fvdb-core#785 · Filter by author swahtz opened on Sep 15, 2026
  1

• Batch the remaining per-member CUDA topology builders

  core libraryoptimizationTopology Operationstriage
  openvdb/fvdb-core#775 · Filter by author swahtz opened on Sep 12, 2026
  4

• TSDF grid allocation is sized by depth pixels that the fusion weights exclude

  externaltriage
  openvdb/fvdb-core#771 · Filter by author lorenzobagnol opened on Sep 11, 2026

• Unscented-transform Gaussian projection has no backward kernel — photometric gradient never reaches means/quats/log_scales for any distorted camera model

  externaltriage
  openvdb/fvdb-core#767 · Filter by author lorenzobagnol opened on Sep 9, 2026

• Depth Loss training using gaussian splat

  externaltriage
  openvdb/fvdb-core#764 · Filter by author Lk13Antkoo opened on Sep 8, 2026

• CI: reuse acquired GPU runners instead of demanding three per PR event

  triage
  openvdb/fvdb-core#761 · Filter by author harrism opened on Sep 3, 2026
  2

• Generative Transposed Convolution

  examplesexternalTopology Operationstriage
  openvdb/fvdb-core#741 · Filter by author lishiyu005 opened on Aug 15, 2026
  3

  

• Unify cu128/cu130 PR build jobs with docker/build_wheel.py

  triage
  openvdb/fvdb-core#736 · Filter by author swahtz opened on Aug 11, 2026

• from_points produces different grids on CPU and CUDA for float16 points at rounding boundaries

  bugcore librarytriage
  openvdb/fvdb-core#725 · Filter by author swahtz opened on Aug 3, 2026
  1

• JaggedTensor.from_data_offsets_and_list_ids accepts malformed joffsets, causing OOB reads in downstream ops

  bugcore librarytriage
  openvdb/fvdb-core#722 · Filter by author swahtz opened on Aug 3, 2026
  

• Evaluate moving PadGrid.cuh functionality to NanoVDB

  NanoVDBTopology Operations
  openvdb/fvdb-core#721 · Filter by author swahtz opened on Aug 3, 2026
  1

• GridEdgeNetwork emits 8 vertices + 12 edges per voxel with no deduplication (~368 B/voxel)

  core libraryoptimizationTopology Operationstriage
  openvdb/fvdb-core#720 · Filter by author swahtz opened on Aug 3, 2026

• Generalize leaf-mask morphology to arbitrary factors: non-power-of-two coarsen/subdivide and general conv strides still pay the coordinate-list + radix-sort path

  NanoVDBoptimizationTopology Operations
  Feature

  openvdb/fvdb-core#716 · Filter by author swahtz opened on Aug 3, 2026

• Move SmoothingMode Enum to a more appropriate module

  triage
  openvdb/fvdb-core#689 · Filter by author swahtz opened on Jul 15, 2026

• marching_cubes: drop redundant batch_idx column from vertex_edge_keys output

  triage
  openvdb/fvdb-core#678 · Filter by author swahtz opened on Jul 2, 2026