ContinuousConvCUDAKernels.h

Go to the documentation of this file.

// ----------------------------------------------------------------------------
// -                        Open3D: www.open3d.org                            -
// ----------------------------------------------------------------------------
// The MIT License (MIT)
//
// Copyright (c) 2018-2021 www.open3d.org
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to deal
// in the Software without restriction, including without limitation the rights
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
// copies of the Software, and to permit persons to whom the Software is
// furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
// IN THE SOFTWARE.
// ----------------------------------------------------------------------------

#pragma once

#include <vector>

#include "open3d/ml/impl/continuous_conv/CoordinateTransformation.cuh"

namespace open3d {
namespace ml {
namespace impl {

template <class TFeat, class TReal, class TIndex>
void FillColumn(const cudaStream_t& stream,
                TFeat* columns,
                int in_channels,
                TIndex begin_idx,
                TIndex end_idx,
                TIndex num_out,
                const TReal* const __restrict__ out_positions,
                TIndex num_inp,
                const TReal* const __restrict__ inp_positions,
                const TFeat* const __restrict__ inp_features,
                const TFeat* const __restrict__ inp_importance,
                size_t neighbors_index_size,
                const TIndex* const __restrict__ neighbors_index,
                const TFeat* const __restrict__ neighbors_importance,
                const int64_t* const __restrict__ neighbors_row_splits,
                const TReal* const __restrict__ extents,
                const TReal* const __restrict__ offsets,
                const std::vector<int>& filter_dims,
                InterpolationMode interpolation,
                CoordinateMapping coordinate_mapping,
                bool align_corners,
                bool individual_extent,
                bool isotropic_extent,
                bool normalize);

template <class TFeat, class TReal, class TIndex>
void FillColumnTranspose(
        const cudaStream_t& stream,
        TFeat* columns,
        int in_channels,
        TIndex begin_idx,
        TIndex end_idx,
        TIndex num_out,
        const TReal* const __restrict__ out_positions,
        TIndex num_inp,
        const TReal* const __restrict__ inp_positions,
        const TFeat* const __restrict__ inp_features,
        const TFeat* const __restrict__ inp_neighbors_importance_sum,
        const int64_t* const __restrict__ inp_neighbors_prefix_sum,
        size_t neighbors_index_size,
        const TIndex* const __restrict__ neighbors_index,
        const TFeat* const __restrict__ neighbors_importance,
        const int64_t* const __restrict__ neighbors_row_splits,
        const TReal* const __restrict__ extents,
        const TReal* const __restrict__ offsets,
        const std::vector<int>& filter_dims,
        InterpolationMode interpolation,
        CoordinateMapping coordinate_mapping,
        bool align_corners,
        bool individual_extent,
        bool isotropic_extent,
        bool normalize);

template <class T>
void MultiplyColumns(const cudaStream_t& stream,
                     size_t rows,
                     size_t cols,
                     T* __restrict__ col_major_matrix,
                     const T* const __restrict__ vector);

template <class T>
void MultiplyAndCopyColumns(const cudaStream_t& stream,
                            size_t rows,
                            size_t cols,
                            T* __restrict__ out_ptr,
                            const T* const __restrict__ col_major_matrix,
                            const T* const __restrict__ vector);

}  // namespace impl
}  // namespace ml
}  // namespace open3d