/************************************************************************
 * Derived from the BSD3-licensed
 * LAPACK routine (version 3.7.0) --
 *     Univ. of Tennessee, Univ. of California Berkeley,
 *     Univ. of Colorado Denver and NAG Ltd..
 *     December 2016
 * Copyright (c) 2020-2021 Advanced Micro Devices, Inc.
 * ***********************************************************************/

#pragma once

#include "auxiliary/rocauxiliary_latrd.hpp"
#include "rocblas.hpp"
#include "roclapack_sytd2_hetd2.hpp"
#include "rocsolver.h"

template <bool BATCHED, typename T>
void rocsolver_sytrd_hetrd_getMemorySize(const rocblas_int n,
                                         const rocblas_int batch_count,
                                         size_t* size_scalars,
                                         size_t* size_work,
                                         size_t* size_norms,
                                         size_t* size_tmptau_W,
                                         size_t* size_workArr)
{
    // if quick return no workspace needed
    if(n == 0 || batch_count == 0)
    {
        *size_scalars = 0;
        *size_work = 0;
        *size_norms = 0;
        *size_tmptau_W = 0;
        *size_workArr = 0;
        return;
    }

    size_t s1 = 0, s2;

    // size required to store temporary matrix W
    if(n > xxTRD_xxTD2_SWITCHSIZE)
    {
        s1 = n * xxTRD_BLOCKSIZE;
        s1 *= sizeof(T) * batch_count;
    }

    // extra requirements to call SYTD2/HETD2
    rocsolver_sytd2_hetd2_getMemorySize<BATCHED, T>(n, batch_count, size_scalars, size_work,
                                                    size_norms, &s2, size_workArr);

    *size_tmptau_W = max(s1, s2);
}

template <typename T, typename S, typename U>
rocblas_status rocsolver_sytrd_hetrd_argCheck(rocblas_handle handle,
                                              const rocblas_fill uplo,
                                              const rocblas_int n,
                                              const rocblas_int lda,
                                              T A,
                                              S D,
                                              S E,
                                              U tau,
                                              const rocblas_int batch_count = 1)
{
    // order is important for unit tests:

    // 1. invalid/non-supported values
    if(uplo != rocblas_fill_upper && uplo != rocblas_fill_lower)
        return rocblas_status_invalid_value;

    // 2. invalid size
    if(n < 0 || lda < n || batch_count < 0)
        return rocblas_status_invalid_size;

    // skip pointer check if querying memory size
    if(rocblas_is_device_memory_size_query(handle))
        return rocblas_status_continue;

    // 3. invalid pointers
    if((n && !A) || (n && !D) || (n && !E) || (n && !tau))
        return rocblas_status_invalid_pointer;

    return rocblas_status_continue;
}

template <bool BATCHED, typename T, typename S, typename U, bool COMPLEX = is_complex<T>>
rocblas_status rocsolver_sytrd_hetrd_template(rocblas_handle handle,
                                              const rocblas_fill uplo,
                                              const rocblas_int n,
                                              U A,
                                              const rocblas_int shiftA,
                                              const rocblas_int lda,
                                              const rocblas_stride strideA,
                                              S* D,
                                              const rocblas_stride strideD,
                                              S* E,
                                              const rocblas_stride strideE,
                                              T* tau,
                                              const rocblas_stride strideP,
                                              const rocblas_int batch_count,
                                              T* scalars,
                                              T* work,
                                              T* norms,
                                              T* tmptau_W,
                                              T** workArr)
{
    ROCSOLVER_ENTER("sytrd_hetrd", "uplo:", uplo, "n:", n, "shiftA:", shiftA, "lda:", lda,
                    "bc:", batch_count);

    // quick return
    if(n == 0 || batch_count == 0)
        return rocblas_status_success;

    hipStream_t stream;
    rocblas_get_stream(handle, &stream);

    rocblas_int k = xxTRD_BLOCKSIZE;
    rocblas_int kk = xxTRD_xxTD2_SWITCHSIZE;

    // if the matrix is too small, use the unblocked variant of the algorithm
    if(n <= kk)
        return rocsolver_sytd2_hetd2_template(handle, uplo, n, A, shiftA, lda, strideA, D, strideD,
                                              E, strideE, tau, strideP, batch_count, scalars, work,
                                              norms, tmptau_W, workArr);

    // everything must be executed with scalars on the host
    rocblas_pointer_mode old_mode;
    rocblas_get_pointer_mode(handle, &old_mode);
    rocblas_set_pointer_mode(handle, rocblas_pointer_mode_host);

    // scalars for rocblas calls
    T minonej = -1; //complex -1
    S one = 1; //real 1

    rocblas_int ldw = n;
    rocblas_stride strideW = n * k;
    rocblas_int j;

    if(uplo == rocblas_fill_lower)
    {
        // reduce the lower part of A
        // main loop running forwards (for each block of columns)
        // when the unreduced part is not large enough, switch to unblocked algorithm
        j = 0;
        while(j < n - kk)
        {
            // reduce columns j:j+k-1
            rocsolver_latrd_template<T>(handle, uplo, n - j, k, A, shiftA + idx2D(j, j, lda), lda,
                                        strideA, (E + j), strideE, (tau + j), strideP, tmptau_W, 0,
                                        ldw, strideW, batch_count, scalars, work, norms, workArr);

            // update unreduced block as a rank-2k update
            // A = A - V*W' - W*V'
            rocblasCall_syr2k_her2k<BATCHED, T>(handle, uplo, rocblas_operation_none, n - j - k, k,
                                                &minonej, A, shiftA + idx2D(j + k, j, lda), lda,
                                                strideA, tmptau_W, idx2D(k, 0, ldw), ldw, strideW,
                                                &one, A, shiftA + idx2D(j + k, j + k, lda), lda,
                                                strideA, batch_count, workArr);

            j += k;
        }

        // reduce last columns of A
        rocsolver_sytd2_hetd2_template<T>(handle, uplo, n - j, A, shiftA + idx2D(j, j, lda), lda,
                                          strideA, (D + j), strideD, (E + j), strideE, (tau + j),
                                          strideP, batch_count, scalars, work, norms, tmptau_W,
                                          workArr);
    }

    else
    {
        // reduce the upper part of A
        // main loop running backwards (for each block of columns)
        // when the unreduced part is not large enough, switch to unblocked algorithm
        j = n - k;
        rocblas_int upkk = n - ((n - kk + k - 1) / k) * k;
        while(j >= upkk)
        {
            // reduce columns j:j+k-1
            rocsolver_latrd_template<T>(handle, uplo, j + k, k, A, shiftA, lda, strideA, E, strideE,
                                        tau, strideP, tmptau_W, 0, ldw, strideW, batch_count,
                                        scalars, work, norms, workArr);

            // update unreduced block as a rank-2k update
            // A = A - V*W' - W*V'
            rocblasCall_syr2k_her2k<BATCHED, T>(handle, uplo, rocblas_operation_none, j, k,
                                                &minonej, A, shiftA + idx2D(0, j, lda), lda,
                                                strideA, tmptau_W, 0, ldw, strideW, &one, A, shiftA,
                                                lda, strideA, batch_count, workArr);
            j -= k;
        }

        // reduce first columns of A
        rocsolver_sytd2_hetd2_template<T>(handle, uplo, upkk, A, shiftA, lda, strideA, D, strideD,
                                          E, strideE, tau, strideP, batch_count, scalars, work,
                                          norms, tmptau_W, workArr);
    }

    // Copy results (set tridiagonal form in A)
    rocblas_int blocks = (n - 1) / BS1 + 1;
    ROCSOLVER_LAUNCH_KERNEL(set_tridiag<T>, dim3(blocks, batch_count), dim3(BS1), 0, stream, uplo,
                            n, A, shiftA, lda, strideA, D, strideD, E, strideE);

    rocblas_set_pointer_mode(handle, old_mode);
    return rocblas_status_success;
}