/* ************************************************************************
 * Copyright (c) 2019-2021 Advanced Micro Devices, Inc.
 * ************************************************************************ */

#include "roclapack_getrf.hpp"

template <typename T, typename U>
rocblas_status rocsolver_getrf_strided_batched_impl(rocblas_handle handle,
                                                    const rocblas_int m,
                                                    const rocblas_int n,
                                                    U A,
                                                    const rocblas_int lda,
                                                    const rocblas_stride strideA,
                                                    rocblas_int* ipiv,
                                                    const rocblas_stride strideP,
                                                    rocblas_int* info,
                                                    const bool pivot,
                                                    const rocblas_int batch_count)
{
    const char* name = (pivot ? "getrf_strided_batched" : "getrf_npvt_strided_batched");
    ROCSOLVER_ENTER_TOP(name, "-m", m, "-n", n, "--lda", lda, "--strideA", strideA, "--strideP",
                        strideP, "--batch_count", batch_count);

    using S = decltype(std::real(T{}));

    if(!handle)
        return rocblas_status_invalid_handle;

    // argument checking
    rocblas_status st
        = rocsolver_getf2_getrf_argCheck(handle, m, n, lda, A, ipiv, info, pivot, batch_count);
    if(st != rocblas_status_continue)
        return st;

    // working with unshifted arrays
    rocblas_int shiftA = 0;
    rocblas_int shiftP = 0;

    // memory workspace sizes:
    // size for constants in rocblas calls
    size_t size_scalars;
    // size of reusable workspace (and for calling TRSM)
    bool optim_mem;
    size_t size_work1, size_work2, size_work3, size_work4;
    // extra requirements for calling GETF2
    size_t size_pivotval, size_pivotidx;
    // size to store info about singularity of each subblock
    size_t size_iinfo, size_iipiv;

    rocsolver_getrf_getMemorySize<false, true, T>(
        m, n, pivot, batch_count, &size_scalars, &size_work1, &size_work2, &size_work3, &size_work4,
        &size_pivotval, &size_pivotidx, &size_iipiv, &size_iinfo, &optim_mem);

    if(rocblas_is_device_memory_size_query(handle))
        return rocblas_set_optimal_device_memory_size(handle, size_scalars, size_work1, size_work2,
                                                      size_work3, size_work4, size_pivotval,
                                                      size_pivotidx, size_iipiv, size_iinfo);

    // memory workspace allocation
    void *scalars, *work1, *work2, *work3, *work4, *pivotval, *pivotidx, *iinfo, *iipiv;
    rocblas_device_malloc mem(handle, size_scalars, size_work1, size_work2, size_work3, size_work4,
                              size_pivotval, size_pivotidx, size_iipiv, size_iinfo);

    if(!mem)
        return rocblas_status_memory_error;

    scalars = mem[0];
    work1 = mem[1];
    work2 = mem[2];
    work3 = mem[3];
    work4 = mem[4];
    pivotval = mem[5];
    pivotidx = mem[6];
    iipiv = mem[7];
    iinfo = mem[8];
    if(size_scalars > 0)
        init_scalars(handle, (T*)scalars);

    // execution
    return rocsolver_getrf_template<false, true, T>(
        handle, m, n, A, shiftA, lda, strideA, ipiv, shiftP, strideP, info, batch_count,
        (T*)scalars, work1, work2, work3, work4, (T*)pivotval, (rocblas_int*)pivotidx,
        (rocblas_int*)iipiv, (rocblas_int*)iinfo, optim_mem, pivot);
}

/*
 * ===========================================================================
 *    C wrapper
 * ===========================================================================
 */

extern "C" {

rocblas_status rocsolver_sgetrf_strided_batched(rocblas_handle handle,
                                                const rocblas_int m,
                                                const rocblas_int n,
                                                float* A,
                                                const rocblas_int lda,
                                                const rocblas_stride strideA,
                                                rocblas_int* ipiv,
                                                const rocblas_stride strideP,
                                                rocblas_int* info,
                                                const rocblas_int batch_count)
{
    return rocsolver_getrf_strided_batched_impl<float>(handle, m, n, A, lda, strideA, ipiv, strideP,
                                                       info, true, batch_count);
}

rocblas_status rocsolver_dgetrf_strided_batched(rocblas_handle handle,
                                                const rocblas_int m,
                                                const rocblas_int n,
                                                double* A,
                                                const rocblas_int lda,
                                                const rocblas_stride strideA,
                                                rocblas_int* ipiv,
                                                const rocblas_stride strideP,
                                                rocblas_int* info,
                                                const rocblas_int batch_count)
{
    return rocsolver_getrf_strided_batched_impl<double>(handle, m, n, A, lda, strideA, ipiv,
                                                        strideP, info, true, batch_count);
}

rocblas_status rocsolver_cgetrf_strided_batched(rocblas_handle handle,
                                                const rocblas_int m,
                                                const rocblas_int n,
                                                rocblas_float_complex* A,
                                                const rocblas_int lda,
                                                const rocblas_stride strideA,
                                                rocblas_int* ipiv,
                                                const rocblas_stride strideP,
                                                rocblas_int* info,
                                                const rocblas_int batch_count)
{
    return rocsolver_getrf_strided_batched_impl<rocblas_float_complex>(
        handle, m, n, A, lda, strideA, ipiv, strideP, info, true, batch_count);
}

rocblas_status rocsolver_zgetrf_strided_batched(rocblas_handle handle,
                                                const rocblas_int m,
                                                const rocblas_int n,
                                                rocblas_double_complex* A,
                                                const rocblas_int lda,
                                                const rocblas_stride strideA,
                                                rocblas_int* ipiv,
                                                const rocblas_stride strideP,
                                                rocblas_int* info,
                                                const rocblas_int batch_count)
{
    return rocsolver_getrf_strided_batched_impl<rocblas_double_complex>(
        handle, m, n, A, lda, strideA, ipiv, strideP, info, true, batch_count);
}

rocblas_status rocsolver_sgetrf_npvt_strided_batched(rocblas_handle handle,
                                                     const rocblas_int m,
                                                     const rocblas_int n,
                                                     float* A,
                                                     const rocblas_int lda,
                                                     const rocblas_stride strideA,
                                                     rocblas_int* info,
                                                     const rocblas_int batch_count)
{
    rocblas_int* ipiv = nullptr;
    return rocsolver_getrf_strided_batched_impl<float>(handle, m, n, A, lda, strideA, ipiv, 0, info,
                                                       false, batch_count);
}

rocblas_status rocsolver_dgetrf_npvt_strided_batched(rocblas_handle handle,
                                                     const rocblas_int m,
                                                     const rocblas_int n,
                                                     double* A,
                                                     const rocblas_int lda,
                                                     const rocblas_stride strideA,
                                                     rocblas_int* info,
                                                     const rocblas_int batch_count)
{
    rocblas_int* ipiv = nullptr;
    return rocsolver_getrf_strided_batched_impl<double>(handle, m, n, A, lda, strideA, ipiv, 0,
                                                        info, false, batch_count);
}

rocblas_status rocsolver_cgetrf_npvt_strided_batched(rocblas_handle handle,
                                                     const rocblas_int m,
                                                     const rocblas_int n,
                                                     rocblas_float_complex* A,
                                                     const rocblas_int lda,
                                                     const rocblas_stride strideA,
                                                     rocblas_int* info,
                                                     const rocblas_int batch_count)
{
    rocblas_int* ipiv = nullptr;
    return rocsolver_getrf_strided_batched_impl<rocblas_float_complex>(
        handle, m, n, A, lda, strideA, ipiv, 0, info, false, batch_count);
}

rocblas_status rocsolver_zgetrf_npvt_strided_batched(rocblas_handle handle,
                                                     const rocblas_int m,
                                                     const rocblas_int n,
                                                     rocblas_double_complex* A,
                                                     const rocblas_int lda,
                                                     const rocblas_stride strideA,
                                                     rocblas_int* info,
                                                     const rocblas_int batch_count)
{
    rocblas_int* ipiv = nullptr;
    return rocsolver_getrf_strided_batched_impl<rocblas_double_complex>(
        handle, m, n, A, lda, strideA, ipiv, 0, info, false, batch_count);
}

} // extern C