Jungfraujoch/image_analysis/bragg_integration/BraggIntegrationEngineGPU.h

// SPDX-FileCopyrightText: 2026 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
// SPDX-License-Identifier: GPL-3.0-only

#pragma once

#include <cstdint>
#include <memory>
#include <vector>

#include "BraggIntegrationEngine.h"
#include "../indexing/CUDAMemHelpers.h"
#include "../../common/PixelMask.h"

// CUDA engine: reproduces BraggIntegrationEngineCPU up to floating-point precision. Each stage is a
// kernel with one CUDA block per reflection cooperating over the small window via shared-memory
// reductions (the natural mapping for thousands of independent, tiny per-spot integrations).
//
// Pipeline (profile modes): reset -> mark_mask -> boxsum -> learn_profile -> build_profiles -> fit
// (the resolution shell is computed inline, so there is no separate shell pass). BoxSum mode stops
// after boxsum (that pass is the BraggIntegrate2D box integrator and the seed of the profile fit).
// The preprocessed image already lives on the device (ImagePreprocessorBufferGPU::getGPUBuffer());
// only the per-frame predicted centres are uploaded.
class BraggIntegrationEngineGPU : public BraggIntegrationEngine {
    std::shared_ptr<CudaStream> stream;
    int threads;
    size_t fit_shared_bytes;
    int rad_w = 0;                 // radial-background window of boxsum, in bins of one pixel
    size_t boxsum_shared_bytes = 0;

    size_t capacity = 0;   // per-reflection device/host arrays hold at least this many reflections

    // Whether d_mask / d_owner may still carry the marks of an earlier image. Run() clears what it
    // marked before it returns when that is cheaper than clearing the frame, and this is then false in
    // the steady state; it is true before the first call, after one that threw part-way through, and
    // whenever the marks covered enough of the frame that clearing all of it was the cheaper choice.
    bool dirty = true;
    size_t mask_box_px = 0;   // pixels one reflection's mark_mask box covers, at the widest aperture

    // --- per-reflection device arrays (grown by EnsureCapacity) ---
    CudaDevicePtr<float>   d_px_x, d_px_y, d_d;
    CudaDevicePtr<uint8_t> d_mark;   // the reflection marks its signal region in d_mask
    CudaDevicePtr<int>     d_cx, d_cy;
    CudaDevicePtr<float>   d_I, d_sigma, d_bkg, d_bkg_var, d_var_bkg, d_obs_x, d_obs_y;
    CudaDevicePtr<float>   d_isum;          // box-sum raw sum, for the radial correction
    CudaDevicePtr<int>     d_ninner, d_rbin, d_kbin;
    // The run's pixel mask, one byte per pixel (PixelMask::GetBinaryMask), shared with the
    // preprocessor's copy. See BraggIntegrationEngineCPU::static_mask.
    std::shared_ptr<CudaDevicePtr<uint8_t>> d_static_mask;
    CudaDevicePtr<uint8_t> d_ok, d_strong, d_has_obs, d_overloaded;

    // --- radial background curvature correction (see BraggIntegrationEngine) ---
    int n_rad = 0;                       // radial bins, 0 when the correction is off
    CudaDevicePtr<unsigned long long> d_rad_sum;   // integer pixel sums, see boxsum
    CudaDevicePtr<float>   d_k_diff;
    CudaDevicePtr<int>     d_rad_cnt;

    // --- fixed-size device arrays ---
    // The learning/fit math is single precision: FP64 is heavily throttled on consumer GPUs and the
    // extraction is Poisson-noise limited, so float reproduces the double CPU path to ~1e-4.
    CudaDevicePtr<uint8_t> d_mask;                       // per-pixel inner-stencil reflection mask
    // Per-pixel (distance, reflection) key naming the nearest predicted centre; allocated only when
    // an overlap treatment is on, so the default path costs no extra device memory.
    CudaDevicePtr<uint32_t> d_owner;
    // Fixed-point (see PROFILE_FIXED): a float atomicAdd here made the profile depend on the order
    // the blocks arrived in, and with it every intensity fitted through it.
    CudaDevicePtr<unsigned long long> d_shell_grid, d_global_grid;  // learned profile accumulators (N_SHELL*GG, GG)
    CudaDevicePtr<float>   d_shell_P, d_global_P;         // normalised profiles (empirical mode)
    CudaDevicePtr<unsigned long long> d_mom;              // learned 2nd moments, 3 per shell + global
    CudaDevicePtr<float>   d_sigma2_r, d_sigma2_t;        // radial/tangential widths, N_SHELL + global
    CudaDevicePtr<int>     d_shell_n, d_global_n;
    CudaDevicePtr<unsigned long long> d_invd2;           // [min,max] inv-d^2 as monotonic bit patterns
    // BraggIntegrationCounts, accumulated on the device so an image costs no transfer; brought back
    // only when Counts() is asked for.
    CudaDevicePtr<unsigned long long> d_counts;

    // --- host staging (copied back once per frame) ---
    // Pinned, like every other engine's staging: a copy out of pageable memory does not return until the
    // driver has staged it through a bounce buffer, so eight of them in a row are eight serialised
    // round-trips rather than eight queued transfers.
    CudaHostPtr<float>   h_px_x, h_px_y, h_d;
    CudaHostPtr<uint8_t> h_mark;
    CudaHostPtr<float>   h_I, h_sigma, h_bkg, h_var_bkg, h_obs_x, h_obs_y;
    CudaHostPtr<uint8_t> h_ok, h_has_obs, h_overloaded;

    void EnsureCapacity(size_t n);

public:
    BraggIntegrationEngineGPU(const DiffractionExperiment &experiment, std::shared_ptr<CudaStream> stream,
                              const PixelMask &mask);
    std::vector<Reflection> Run(const ImagePreprocessorBuffer &image,
                                const std::vector<Reflection> &predicted, size_t npredicted,
                                int64_t image_number) override;

    // Brings the two device counters back before answering. Synchronises the stream, so ask once a
    // pass rather than once an image.
    [[nodiscard]] BraggIntegrationCounts Counts() const override;
};