// SPDX-FileCopyrightText: 2025 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
// SPDX-License-Identifier: GPL-3.0-only

#ifndef JFJOCH_IMAGEANALYSISGPU_H
#define JFJOCH_IMAGEANALYSISGPU_H

#include <vector>

#include "SpotFindingSettings.h"
#include "ImageSpotFinder.h"
#include "../indexing/CUDAMemHelpers.h"

class ImageSpotFinderGPU : public ImageSpotFinder {
    CudaStream stream;

    CudaDevicePtr<int32_t> gpu_in;
    CudaDevicePtr<uint32_t> gpu_out_0;
    CudaDevicePtr<uint32_t> gpu_out_1;

    CudaRegisteredVector<int32_t> input_buffer_reg;
    CudaRegisteredVector<uint32_t> output_buffer_reg;

    const int numberOfCudaThreads = 128;    // #threads per block that should work well for Nvidia L4
    const int numberOfWaves = 32;           // #waves that should work well for Nvidia L4
    const int windowSizeLimit = 32;         // limit on the window size (2nby+1, 2nbx+1) to prevent shared memory problems
public:
    ImageSpotFinderGPU(int32_t width, int32_t height);
    ~ImageSpotFinderGPU() override = default;

    std::vector<DiffractionSpot> Run(const SpotFindingSettings &settings, const std::vector<bool> &res_mask) override;
};

#endif //JFJOCH_IMAGEANALYSISGPU_H