src/cpu/kernels/softmax/generic/neon/impl.cpp

/*
 * Copyright (c) 2021-2023 Arm Limited.
 *
 * SPDX-License-Identifier: MIT
 *
 * 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.
 */
#include "src/cpu/kernels/softmax/generic/neon/impl.h"

#include "support/SaturateCast.h"

namespace arm_compute
{
namespace cpu
{
template <typename T, bool IS_LOG>
void neon_softmax_quantized(const ITensor *in, void *const tmp, ITensor *out, float beta, const Window &window)
{
    static_assert(std::is_same<T, qasymm8_t>::value || std::is_same<T, qasymm8_signed_t>::value,
                  "quantized type should be either qasymm8_t or qasymm8_signed_t.");

    const int input_width = in->info()->valid_region().shape.x();

    const float       scale_beta     = -beta * in->info()->quantization_info().uniform().scale;
    const float32x4_t scale_beta_vec = vdupq_n_f32(scale_beta);

    Iterator in_it(in, window);
    Iterator out_it(out, window);

    constexpr int vec_size = 16;

#ifndef __aarch64__
    const int sum_stages = log2(vec_size >> 1);
#endif // __aarch64__

    using ExactTagType = typename wrapper::traits::neon_bitvector_tag_t<T, wrapper::traits::BitWidth::W128>;

    execute_window_loop(
        window,
        [&](const Coordinates &)
        {
            /* Get pointers */
            const T *in_ptr  = reinterpret_cast<const T *>(in_it.ptr());
            T       *out_ptr = reinterpret_cast<T *>(out_it.ptr());
            float   *tmp_ptr = reinterpret_cast<float *>(tmp);

            T max_val;

            /* Compute Max */
            {
                // Init max value
                auto vec_max = wrapper::vdup_n(support::cpp11::lowest<T>(), ExactTagType{});
                int  x       = 0;

                for (; x <= (input_width - vec_size); x += vec_size)
                {
                    const auto current_value = wrapper::vloadq(in_ptr + x);
                    vec_max                  = wrapper::vmax(vec_max, current_value);
                }

#ifdef __aarch64__
                max_val = wrapper::vmaxv(vec_max);
#else  // __aarch64__
                auto carry_max = wrapper::vpmax(wrapper::vgethigh(vec_max), wrapper::vgetlow(vec_max));

                for (int i = 0; i < sum_stages; ++i)
                {
                    carry_max = wrapper::vpmax(carry_max, carry_max);
                }

                max_val      = wrapper::vgetlane(carry_max, 0);
#endif // __aarch64__

                // Compute left-over elements
                for (; x < input_width; ++x)
                {
                    max_val = std::max(*(in_ptr + x), max_val);
                }
            } // Compute Max

            float sum_transformed{};

            /* Compute exponentials and sum */
            {
                /* Get max value */
                const auto vec_max = wrapper::vdup_n(max_val, wrapper::traits::vector_128_tag{});

                /* Init sum to zero */
                float32x4x4_t vec_sum = {
                    vdupq_n_f32(0.f),
                    vdupq_n_f32(0.f),
                    vdupq_n_f32(0.f),
                    vdupq_n_f32(0.f),
                };

                /* Loop over row and compute exponentials and sum */
                int x = 0;
                for (; x <= (input_width - vec_size); x += vec_size)
                {
                    auto vec_elements              = wrapper::vloadq(in_ptr + x);
                    vec_elements                   = wrapper::vqsub(vec_max, vec_elements);
                    float32x4x4_t vec_elements_flt = convert_int_to_float<float32x4x4_t>(vec_elements);

                    if (IS_LOG)
                    {
                        vec_elements_flt.val[0] = vmulq_f32(vec_elements_flt.val[0], scale_beta_vec);
                        vec_elements_flt.val[1] = vmulq_f32(vec_elements_flt.val[1], scale_beta_vec);
                        vec_elements_flt.val[2] = vmulq_f32(vec_elements_flt.val[2], scale_beta_vec);
                        vec_elements_flt.val[3] = vmulq_f32(vec_elements_flt.val[3], scale_beta_vec);
                        vec_sum.val[0]          = vaddq_f32(vec_sum.val[0], vexpq_f32(vec_elements_flt.val[0]));
                        vec_sum.val[1]          = vaddq_f32(vec_sum.val[1], vexpq_f32(vec_elements_flt.val[1]));
                        vec_sum.val[2]          = vaddq_f32(vec_sum.val[2], vexpq_f32(vec_elements_flt.val[2]));
                        vec_sum.val[3]          = vaddq_f32(vec_sum.val[3], vexpq_f32(vec_elements_flt.val[3]));
                    }
                    else
                    {
                        vec_elements_flt.val[0] = vexpq_f32(vmulq_f32(vec_elements_flt.val[0], scale_beta_vec));
                        vec_elements_flt.val[1] = vexpq_f32(vmulq_f32(vec_elements_flt.val[1], scale_beta_vec));
                        vec_elements_flt.val[2] = vexpq_f32(vmulq_f32(vec_elements_flt.val[2], scale_beta_vec));
                        vec_elements_flt.val[3] = vexpq_f32(vmulq_f32(vec_elements_flt.val[3], scale_beta_vec));
                        vec_sum.val[0]          = vaddq_f32(vec_sum.val[0], vec_elements_flt.val[0]);
                        vec_sum.val[1]          = vaddq_f32(vec_sum.val[1], vec_elements_flt.val[1]);
                        vec_sum.val[2]          = vaddq_f32(vec_sum.val[2], vec_elements_flt.val[2]);
                        vec_sum.val[3]          = vaddq_f32(vec_sum.val[3], vec_elements_flt.val[3]);
                    }

                    vst4q_f32(tmp_ptr + x, vec_elements_flt);
                }

                /* Reduce sum */
                const float32x4_t sum_16_byte =
                    vaddq_f32(vaddq_f32(vec_sum.val[0], vec_sum.val[1]), vaddq_f32(vec_sum.val[2], vec_sum.val[3]));

                float sum;

#ifdef __aarch64__
                sum = wrapper::vaddv(sum_16_byte);
#else  // __aarch64__
                auto sum_res = vpadd_f32(vget_high_f32(sum_16_byte), vget_low_f32(sum_16_byte));
                sum_res      = vpadd_f32(sum_res, sum_res);
                sum          = wrapper::vgetlane(sum_res, 0);
#endif // __aarch64__

                /* Run remaining elements */
                for (; x < input_width; ++x)
                {
                    float element{};
                    if (IS_LOG)
                    {
                        element = (max_val - in_ptr[x]) * scale_beta;
                        sum += std::exp(element);
                    }
                    else
                    {
                        element = std::exp((max_val - in_ptr[x]) * scale_beta);
                        sum += element;
                    }

                    tmp_ptr[x] = element;
                }

                if (!IS_LOG)
                {
                    sum_transformed = 256.f / sum;
                }
                else
                {
                    sum_transformed = std::log(sum);
                }
            } // Compute exponentials and sum

            /* Normalize exponentials */
            {
                constexpr bool is_qasymm8_signed = std::is_same<T, qasymm8_signed_t>::value;

                const float32x4_t sum_vec = vdupq_n_f32(sum_transformed);

                /* Loop over row and compute softmax */
                int x = 0;
                for (; x <= (input_width - vec_size); x += vec_size)
                {
                    using int_vec_type   = wrapper::traits::neon_vector_t<T, 16>;
                    float32x4x4_t vec_in = vld4q_f32(tmp_ptr + x);
                    int_vec_type  normalized_value{};
                    if (IS_LOG)
                    {
                        const float32x4x4_t sub = {
                            vsubq_f32(vec_in.val[0], sum_vec),
                            vsubq_f32(vec_in.val[1], sum_vec),
                            vsubq_f32(vec_in.val[2], sum_vec),
                            vsubq_f32(vec_in.val[3], sum_vec),
                        };
                        normalized_value = convert_float_to_int<float32x4x4_t, int_vec_type>(sub);
                    }
                    else
                    {
                        float32x4x4_t mul = {
                            vmulq_f32(vec_in.val[0], sum_vec),
                            vmulq_f32(vec_in.val[1], sum_vec),
                            vmulq_f32(vec_in.val[2], sum_vec),
                            vmulq_f32(vec_in.val[3], sum_vec),
                        };

                        if (is_qasymm8_signed)
                        {
                            const auto offset_vec = wrapper::vdup_n(128.f, wrapper::traits::vector_128_tag{});
                            mul.val[0]            = wrapper::vsub(mul.val[0], offset_vec);
                            mul.val[1]            = wrapper::vsub(mul.val[1], offset_vec);
                            mul.val[2]            = wrapper::vsub(mul.val[2], offset_vec);
                            mul.val[3]            = wrapper::vsub(mul.val[3], offset_vec);
                        }

                        normalized_value = convert_float_to_int<float32x4x4_t, int_vec_type>(mul);
                    }
                    wrapper::vstore(out_ptr + x, normalized_value);
                }
                /* Run remaining elements */
                for (; x < input_width; ++x)
                {
                    if (IS_LOG)
                    {
                        out_ptr[x] = utils::cast::saturate_cast<T>(tmp_ptr[x] - sum_transformed);
                    }
                    else
                    {
                        out_ptr[x] = utils::cast::saturate_cast<T>((tmp_ptr[x] * sum_transformed) -
                                                                   (is_qasymm8_signed ? 128.f : 0));
                    }
                }
            } // Normalize exponentials
        },
        in_it, out_it);
}

template void neon_softmax_quantized<qasymm8_signed_t, true>(
    const ITensor *in, void *const tmp, ITensor *out, float beta, const Window &window);

template void neon_softmax_quantized<qasymm8_signed_t, false>(
    const ITensor *in, void *const tmp, ITensor *out, float beta, const Window &window);

template void neon_softmax_quantized<qasymm8_t, true>(
    const ITensor *in, void *const tmp, ITensor *out, float beta, const Window &window);

template void neon_softmax_quantized<qasymm8_t, false>(
    const ITensor *in, void *const tmp, ITensor *out, float beta, const Window &window);
} // namespace cpu
} // namespace arm_compute