Skip to content

Update nanovdb to version 32.9.0 and refine grid type checks #475

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
swahtz merged 2 commits into from
Feb 24, 2026
+36 −116
Merged

Update nanovdb to version 32.9.0 and refine grid type checks #475

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b24bdf5
Update nanovdb to version 32.9.0 and refine grid type checks
swahtz Feb 24, 2026
917bdc8
docs fix
swahtz Feb 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion src/cmake/get_nanovdb.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@
CPMAddPackage(
NAME nanovdb
GITHUB_REPOSITORY AcademySoftwareFoundation/openvdb
GIT_TAG 5f0432b3387c169212a009ddaa05fdd703016549
GIT_TAG cfba6dff785629c9f02583fde6e38bf845b3509b
SOURCE_SUBDIR nanovdb/nanovdb
DOWNLOAD_ONLY YES
)
Expand Down
5 changes: 2 additions & 3 deletions src/fvdb/detail/GridBatchImpl.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -869,9 +869,8 @@ GridBatchImpl::setGrid(nanovdb::GridHandle<TorchDeviceBuffer> &&gridHdl,
gridHdl.gridCount());
TORCH_CHECK(voxelOrigins.size() == gridHdl.gridCount(),
"Voxel origins must be the same size as the number of grids");
TORCH_CHECK((gridHdl.gridType(0) == nanovdb::GridType::OnIndex) ||
(gridHdl.gridType(0) == nanovdb::GridType::OnIndexMask),
"GridBatchImpl only supports ValueOnIndex and ValueOnIndexMask grids");
TORCH_CHECK(gridHdl.gridType(0) == nanovdb::GridType::OnIndex,
"GridBatchImpl only supports ValueOnIndex grids");

// Reallocate GridMetadata
mBatchSize = gridHdl.gridCount();
Expand Down
141 changes: 33 additions & 108 deletions src/fvdb/detail/io/LoadNanovdb.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,12 +127,10 @@ isFvdbBlindData(const nanovdb::GridBlindMetaData &blindMetadata) {
return std::make_tuple(true, std::optional<torch::Dtype>(blindDtype));
}

/// @brief Copy a source index grid (ValueIndex(Mask) or ValueOnIndex(Mask)) to a
/// @brief Copy a source index grid (ValueIndex or ValueOnIndex) to a
/// nanovdb::GridHandle<TorchDeviceBuffer>.
/// If the source type is ValueIndex or ValueIndex mask it will be set to ValueOnIndex or
/// ValueOnIndexMask respectively.
/// @tparam SourceGridT The type of the source grid (must be a nanovdb::ValueIndex or
/// nanovdb::ValueIndexMask)
/// If the source type is ValueIndex it will be set to ValueOnIndex.
/// @tparam SourceGridT The type of the source grid (must be a nanovdb::ValueIndex)
/// @tparam TargetGridT The type of the target grid (must be a form of index grid)
/// @param sourceGrid A host pointer to the source grid to copy
/// @return A handle to the copied grid
Expand All @@ -141,27 +139,15 @@ nanovdb::GridHandle<TorchDeviceBuffer>
copyIndexGridToHandle(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
constexpr bool isSrcValueOnIndex =
nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndex>::value;
constexpr bool isSrcValueOnIndexMask =
nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndexMask>::value;
constexpr bool isSrcValueIndex =
nanovdb::util::is_same<SourceGridT, nanovdb::ValueIndex>::value;
constexpr bool isSrcValueIndexMask =
nanovdb::util::is_same<SourceGridT, nanovdb::ValueIndexMask>::value;
constexpr bool isTgtValueOnIndex =
nanovdb::util::is_same<TargetGridT, nanovdb::ValueOnIndex>::value;
constexpr bool isTgtValueOnIndexMask =
nanovdb::util::is_same<TargetGridT, nanovdb::ValueOnIndexMask>::value;

static_assert(isSrcValueOnIndex || isSrcValueOnIndexMask || isSrcValueIndex ||
isSrcValueIndexMask,
static_assert(isSrcValueOnIndex || isSrcValueIndex,
"Bad source type in copyIndexGridToHandle must be an Index grid type.");
static_assert(
isTgtValueOnIndex || isTgtValueOnIndexMask,
"Bad target type in copyIndexGridToHandle must be ValueOnIndex or ValueOnIndexMask.");
static_assert(
(isTgtValueOnIndex && (isSrcValueIndex || isSrcValueOnIndex)) ||
(isTgtValueOnIndexMask && (isSrcValueIndexMask || isSrcValueOnIndexMask)),
"Bad target grid type for given source grid type in copyIndexGridToHandle. If source is a masked grid, then target must also be a masked grid.");
static_assert(isTgtValueOnIndex,
"Bad target type in copyIndexGridToHandle must be ValueOnIndex.");

const ptrdiff_t gridSize =
sourceGrid->blindDataCount() > 0
Expand All @@ -177,12 +163,10 @@ copyIndexGridToHandle(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
return nanovdb::GridHandle<TorchDeviceBuffer>(std::move(buf));
}

/// @brief Load a nanovdb ValueOnIndex or ValueOnIndexMask grid with tensor blind metatada
/// (GridClass = TensorGrid) into
/// @brief Load a nanovdb ValueOnIndex grid with tensor blind metatada (GridClass = TensorGrid) into
/// an index grid of the same type stored in a TorchDeviceBuffer) and a torch tensor of data
/// (i.e. the standard grid format for FVDB).
/// @tparam SourceGridT The type of the source grid (must be a nanovdb::ValueOnIndex or
/// nanovdb::ValueOnIndexMask)
/// @tparam SourceGridT The type of the source grid (must be a nanovdb::ValueOnIndex)
/// @tparam TargetGridT The type of the target grid (must be a form of index grid)
/// @param sourceGrid A host pointer to the source grid to load
/// @return A tuple containing the index grid, the name of the grid, the tensor of data, the voxel
Expand All @@ -194,14 +178,10 @@ std::tuple<nanovdb::GridHandle<TorchDeviceBuffer>,
nanovdb::Vec3d,
nanovdb::Vec3d>
nanovdbTensorGridToFVDBGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
static_assert(
nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndex>::value ||
nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndexMask>::value,
"Bad source grid type in nanovdbTensorGridToFVDBGrid. Must be ValueOnIndex or ValueOnIndexMask.");
static_assert(
nanovdb::util::is_same<TargetGridT, nanovdb::ValueOnIndex>::value ||
nanovdb::util::is_same<TargetGridT, nanovdb::ValueOnIndexMask>::value,
"Bad target grid type in nanovdbTensorGridToFVDBGrid. Must be ValueOnIndex or ValueOnIndexMask.");
static_assert(nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndex>::value,
"Bad source grid type in nanovdbTensorGridToFVDBGrid. Must be ValueOnIndex.");
static_assert(nanovdb::util::is_same<TargetGridT, nanovdb::ValueOnIndex>::value,
"Bad target grid type in nanovdbTensorGridToFVDBGrid. Must be ValueOnIndex.");
static_assert(
nanovdb::util::is_same<SourceGridT, TargetGridT>::value,
"Mismatched source and target grid types in nanovdbTensorGridToFVDBGrid. They must be identical.");
Expand Down Expand Up @@ -273,13 +253,11 @@ nanovdbTensorGridToFVDBGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
return std::make_tuple(std::move(retHandle), name, retData, voxSize, voxOrigin);
}

/// @brief Load a nanovdb index grid (ValueOnIndex(Mask) or ValueIndex(Mask)) into an ValueOnIndex
/// or ValueIndex grid
/// (stored in a TorchDeviceBuffer) and an empty tensor of data (i.e. the standard grid
/// format for FVDB).
/// @brief Load a nanovdb index grid (ValueOnIndex or ValueIndex) into an ValueOnIndex or
/// ValueIndex grid (stored in a TorchDeviceBuffer) and an empty tensor of data (i.e. the
/// standard grid format for FVDB).
/// @tparam SourceGridT The type of the source grid (must not be an index grid)
/// @tparam TargetGridT The type of the target grid (must be a nanovdb::ValueOnIndex or
/// nanovdb::ValueOnIndexMask)
/// @tparam TargetGridT The type of the target grid (must be a nanovdb::ValueOnIndex)
/// @param sourceGrid A host pointer to the source grid to load
/// @return A tuple containing the index grid, the name of the grid, the empty tensor of data, the
/// voxel size, and the voxel origin
Expand All @@ -303,8 +281,7 @@ nanovdbIndexGridToFVDBGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
/// (stored in a TorchDeviceBuffer) and a tensor of data (i.e. the standard grid format for
/// FVDB).
/// @tparam SourceGridT The type of the source grid (must not be an index grid)
/// @tparam TargetGridT The type of the target grid (must be a nanovdb::ValueOnIndex or
/// nanovdb::ValueOnIndexMask)
/// @tparam TargetGridT The type of the target grid (must be a nanovdb::ValueOnIndex)
/// @tparam ScalarType The scalar type of data stored in the source grid
/// @tparam DataDim The dimension of the data stored in the source grid
/// @param sourceGrid A host pointer to the source grid to load
Expand All @@ -322,9 +299,7 @@ nanovdbGridToFvdbGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {

static_assert(
!nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndex>::value &&
!nanovdb::util::is_same<SourceGridT, nanovdb::ValueOnIndexMask>::value &&
!nanovdb::util::is_same<SourceGridT, nanovdb::ValueIndex>::value &&
!nanovdb::util::is_same<SourceGridT, nanovdb::ValueIndexMask>::value,
!nanovdb::util::is_same<SourceGridT, nanovdb::ValueIndex>::value,
"Bad source type in nanovdbGridToIndexGridAndData must NOT be an Index grid type.");

// Create the index grid for the loaded grid
Expand All @@ -343,8 +318,7 @@ nanovdbGridToFvdbGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
TORCH_CHECK(outGrid != nullptr, "Internal error: failed to get outGrid.");
TORCH_CHECK(outGrid->gridClass() == nanovdb::GridClass::IndexGrid,
"Internal error: outGrid is not an index grid.");
TORCH_CHECK(outGrid->gridType() == nanovdb::GridType::OnIndex ||
outGrid->gridType() == nanovdb::GridType::OnIndexMask,
TORCH_CHECK(outGrid->gridType() == nanovdb::GridType::OnIndex,
"Internal error: outGrid is not an index grid.");

// Load data at the voxels into a tensor
Expand Down Expand Up @@ -414,20 +388,21 @@ nanovdbGridToFvdbGrid(const nanovdb::NanoGrid<SourceGridT> *sourceGrid) {
}

/// @brief Load a single nanovdb grid in a nanovdb::GridHandle<nanovdb::HostBuffer> into an
/// ValueOnIndex or ValueOnIndexMask grid
/// ValueOnIndex grid
/// stored in a nanovdb::GridHandle<TorchDeviceBuffer> as well as torch::Tensor encoding the
/// data at the voxels (i.e. the standard format for FVDB). There are 3 cases:
/// 1. The input grid has scalar or vector values at the leaves:
/// - Load a ValueOnIndex grid and torch::Tensor of values
/// 2. The input grid is a ValueOnIndex or ValueOnIndexMask and has its grid class set to
/// 2. The input grid is a ValueOnIndex and has its grid class set to
/// TensorGrid:
/// - Load a matching ValueOnIndex or ValueOnIndexMask grid and torch::Tensor of values
/// - Load a matching ValueOnIndex grid and torch::Tensor of values
/// corresponding to
/// the blind data (if it is present)
/// 3. The input grid is an index grid (ValueIndex(Mask) or ValueOnIndex(Mask)) but doesn't
/// 3. The input grid is an index grid (ValueIndex or ValueOnIndex) but doesn't
/// have a TensorGrid class set:
/// - Load a ValueOnIndex or ValueOnIndexMask grid (depending if the input type has a
/// mask or not) and an empty torch::Tensor of values
/// - Load a ValueOnIndex grid and an empty torch::Tensor of values and an empty
/// torch::Tensor of values
///
///
/// @param handle The grid handle to read from
/// @param gridId The index of the grid in the handle to read
Expand All @@ -442,20 +417,12 @@ std::tuple<nanovdb::GridHandle<TorchDeviceBuffer>,
loadOneGrid(const nanovdb::GridHandle<nanovdb::HostBuffer> &handle, uint32_t gridId, uint32_t bi) {
if (handle.gridMetaData()->gridClass() == nanovdb::GridClass::TensorGrid) {
TORCH_CHECK(
handle.gridType() == nanovdb::GridType::OnIndex ||
handle.gridType() == nanovdb::GridType::OnIndexMask,
handle.gridType() == nanovdb::GridType::OnIndex,
"Invalid grid type: Tensor grids which are not saved with fVDB are not yet supported.");
if (handle.gridType() == nanovdb::GridType::OnIndex) {
const nanovdb::NanoGrid<nanovdb::ValueOnIndex> *sourceGrid =
getGrid<nanovdb::ValueOnIndex>(handle, gridId, bi);
return nanovdbTensorGridToFVDBGrid<nanovdb::ValueOnIndex, nanovdb::ValueOnIndex>(
sourceGrid);
} else if (handle.gridType() == nanovdb::GridType::OnIndexMask) {
const nanovdb::NanoGrid<nanovdb::ValueOnIndexMask> *sourceGrid =
getGrid<nanovdb::ValueOnIndexMask>(handle, gridId, bi);
return nanovdbTensorGridToFVDBGrid<nanovdb::ValueOnIndexMask,
nanovdb::ValueOnIndexMask>(sourceGrid);
}
const nanovdb::NanoGrid<nanovdb::ValueOnIndex> *sourceGrid =
getGrid<nanovdb::ValueOnIndex>(handle, gridId, bi);
return nanovdbTensorGridToFVDBGrid<nanovdb::ValueOnIndex, nanovdb::ValueOnIndex>(
sourceGrid);
}

switch (handle.gridType()) {
Expand Down Expand Up @@ -519,23 +486,11 @@ loadOneGrid(const nanovdb::GridHandle<nanovdb::HostBuffer> &handle, uint32_t gri
getGrid<nanovdb::ValueIndex>(handle, gridId, bi);
return nanovdbIndexGridToFVDBGrid<nanovdb::ValueIndex, nanovdb::ValueOnIndex>(sourceGrid);
}
case nanovdb::GridType::IndexMask: {
const nanovdb::NanoGrid<nanovdb::ValueIndexMask> *sourceGrid =
getGrid<nanovdb::ValueIndexMask>(handle, gridId, bi);
return nanovdbIndexGridToFVDBGrid<nanovdb::ValueIndexMask, nanovdb::ValueOnIndexMask>(
sourceGrid);
}
case nanovdb::GridType::OnIndex: {
const nanovdb::NanoGrid<nanovdb::ValueOnIndex> *sourceGrid =
getGrid<nanovdb::ValueOnIndex>(handle, gridId, bi);
return nanovdbIndexGridToFVDBGrid<nanovdb::ValueOnIndex, nanovdb::ValueOnIndex>(sourceGrid);
}
case nanovdb::GridType::OnIndexMask: {
const nanovdb::NanoGrid<nanovdb::ValueOnIndexMask> *sourceGrid =
getGrid<nanovdb::ValueOnIndexMask>(handle, gridId, bi);
return nanovdbIndexGridToFVDBGrid<nanovdb::ValueOnIndexMask, nanovdb::ValueOnIndexMask>(
sourceGrid);
}
default:
// Unhandled cases include: Int16, UInt32, Fp4, Fp8, FpN
char gridTypeStr[nanovdb::strlen<nanovdb::GridType>()];
Expand All @@ -552,8 +507,7 @@ fromNVDB(nanovdb::GridHandle<nanovdb::HostBuffer> &handle,
std::vector<nanovdb::GridHandle<TorchDeviceBuffer>> grids;
std::vector<nanovdb::Vec3d> voxSizes, voxOrigins;
std::vector<std::string> names;
uint32_t bi = 0;
nanovdb::GridType lastGridType = nanovdb::GridType::Unknown;
uint32_t bi = 0;
for (size_t gridId = 0; gridId < handle.gridCount(); gridId += 1) {
auto gridData = loadOneGrid(handle, gridId, bi);
grids.push_back(std::move(std::get<0>(gridData)));
Expand All @@ -562,20 +516,6 @@ fromNVDB(nanovdb::GridHandle<nanovdb::HostBuffer> &handle,
voxSizes.push_back(std::move(std::get<3>(gridData)));
voxOrigins.push_back(std::move(std::get<4>(gridData)));

// FVDB grid batches all share the same mutability. i.e. a grid batch consists of all
// ValueOnIndex (immutable) grids or all ValueOnIndexMask (mutable) grids.
// In all but two cases, we load a ValueOnIndex grid and a tensor of data:
// 1. When the user saved a mutable Tensor grid with save
// 2. When the user loaded a batch with a ValueOnIndexMask grid
// If the file the list of grids the user loaded contains a mix of ValueOnIndex and
// ValueOnIndexMask grids, then it's unclear what to do, so throw an exception.
if (bi > 0) {
TORCH_CHECK(
lastGridType == grids.back().gridData()->mGridType,
"All grids in a batch must have the same mutability (i.e. all ValueOnIndex or all ValueOnIndexMask).");
}
lastGridType = grids.back().gridData()->mGridType;

bi += 1;
}

Expand Down Expand Up @@ -610,8 +550,7 @@ fromNVDB(const std::vector<nanovdb::GridHandle<nanovdb::HostBuffer>> &handles,
std::vector<nanovdb::GridHandle<TorchDeviceBuffer>> grids;
std::vector<nanovdb::Vec3d> voxSizes, voxOrigins;
std::vector<std::string> names;
uint32_t bi = 0;
nanovdb::GridType lastGridType = nanovdb::GridType::Unknown;
uint32_t bi = 0;
for (const auto &handle: handles) {
for (size_t gridId = 0; gridId < handle.gridCount(); gridId += 1) {
auto gridData = loadOneGrid(handle, gridId, bi);
Expand All @@ -621,20 +560,6 @@ fromNVDB(const std::vector<nanovdb::GridHandle<nanovdb::HostBuffer>> &handles,
voxSizes.push_back(std::move(std::get<3>(gridData)));
voxOrigins.push_back(std::move(std::get<4>(gridData)));

// FVDB grid batches all share the same mutability. i.e. a grid batch consists of all
// ValueOnIndex (immutable) grids or all ValueOnIndexMask (mutable) grids.
// In all but two cases, we load a ValueOnIndex grid and a tensor of data:
// 1. When the user saved a mutable Tensor grid with save
// 2. When the user loaded a batch with a ValueOnIndexMask grid
// If the file the list of grids the user loaded contains a mix of ValueOnIndex and
// ValueOnIndexMask grids, then it's unclear what to do, so throw an exception.
if (bi > 0) {
TORCH_CHECK(
lastGridType == grids.back().gridData()->mGridType,
"All grids in a batch must have the same mutability (i.e. all ValueOnIndex or all ValueOnIndexMask).");
}
lastGridType = grids.back().gridData()->mGridType;

bi += 1;
}
}
Expand Down
4 changes: 0 additions & 4 deletions src/fvdb/detail/utils/nanovdb/ActiveVoxelIterator.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,10 +21,6 @@ template <typename TreeT> struct ActiveVoxelIteratorDataTypeExtractor {
template <> struct ActiveVoxelIteratorDataTypeExtractor<nanovdb::OnIndexTree> {
using DataType = int64_t;
};
template <>
struct ActiveVoxelIteratorDataTypeExtractor<nanovdb::NanoTree<nanovdb::ValueOnIndexMask>> {
using DataType = int64_t;
};

/*
* Const iterator over voxels in a nanovdb Index grid
Expand Down
Loading