xref: /external/XNNPACK/src/f32-ibilinear-chw/gen/wasmsimd-p8.c

// Auto-generated file. Do not edit!
//   Template: src/f32-ibilinear-chw/wasmsimd.c.in
//   Generator: tools/xngen
//
// Copyright 2020 Google LLC
//
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree.

#include <assert.h>

#include <wasm_simd128.h>

#include <xnnpack/ibilinear.h>


void xnn_f32_ibilinear_chw_ukernel__wasmsimd_p8(
    size_t output_pixels,
    size_t channels,
    const float**restrict input,
    size_t input_offset,
    const float*restrict weights,
    float*restrict output,
    size_t input_increment) XNN_OOB_READS
{
  assert(output_pixels != 0);
  assert(channels != 0);
  assert(input_increment % sizeof(float) == 0);

  do {
    const float** i = input;
    const float* w = weights;
    size_t p = output_pixels;
    for (; p >= 8; p -= 8) {
      const float* itl0 = (const float*) ((uintptr_t) i[0] + input_offset);
      const float* ibl0 = (const float*) ((uintptr_t) i[1] + input_offset);
      const float* itl1 = (const float*) ((uintptr_t) i[2] + input_offset);
      const float* ibl1 = (const float*) ((uintptr_t) i[3] + input_offset);
      const float* itl2 = (const float*) ((uintptr_t) i[4] + input_offset);
      const float* ibl2 = (const float*) ((uintptr_t) i[5] + input_offset);
      const float* itl3 = (const float*) ((uintptr_t) i[6] + input_offset);
      const float* ibl3 = (const float*) ((uintptr_t) i[7] + input_offset);
      const float* itl4 = (const float*) ((uintptr_t) i[8] + input_offset);
      const float* ibl4 = (const float*) ((uintptr_t) i[9] + input_offset);
      const float* itl5 = (const float*) ((uintptr_t) i[10] + input_offset);
      const float* ibl5 = (const float*) ((uintptr_t) i[11] + input_offset);
      const float* itl6 = (const float*) ((uintptr_t) i[12] + input_offset);
      const float* ibl6 = (const float*) ((uintptr_t) i[13] + input_offset);
      const float* itl7 = (const float*) ((uintptr_t) i[14] + input_offset);
      const float* ibl7 = (const float*) ((uintptr_t) i[15] + input_offset);
      i += 2 * 8;

      const v128_t vw0123p0 = wasm_v128_load(w + 0);
      const v128_t vw0123p1 = wasm_v128_load(w + 4);
      const v128_t vw4567p0 = wasm_v128_load(w + 8);
      const v128_t vw4567p1 = wasm_v128_load(w + 12);
      w += 2 * 8;

      const v128_t vtltr0 = wasm_v128_load64_splat(itl0);
      const v128_t vblbr0 = wasm_v128_load64_splat(ibl0);
      const double vtltr1 = *((const double*) itl1);
      const double vblbr1 = *((const double*) ibl1);
      const v128_t vtltr2 = wasm_v128_load64_splat(itl2);
      const v128_t vblbr2 = wasm_v128_load64_splat(ibl2);
      const double vtltr3 = *((const double*) itl3);
      const double vblbr3 = *((const double*) ibl3);
      const v128_t vtltr4 = wasm_v128_load64_splat(itl4);
      const v128_t vblbr4 = wasm_v128_load64_splat(ibl4);
      const double vtltr5 = *((const double*) itl5);
      const double vblbr5 = *((const double*) ibl5);
      const v128_t vtltr6 = wasm_v128_load64_splat(itl6);
      const v128_t vblbr6 = wasm_v128_load64_splat(ibl6);
      const double vtltr7 = *((const double*) itl7);
      const double vblbr7 = *((const double*) ibl7);

      const v128_t valphah0123 = wasm_v32x4_shuffle(vw0123p0, vw0123p1, 0, 2, 4, 6);
      const v128_t valphav0123 = wasm_v32x4_shuffle(vw0123p0, vw0123p1, 1, 3, 5, 7);
      const v128_t valphah4567 = wasm_v32x4_shuffle(vw4567p0, vw4567p1, 0, 2, 4, 6);
      const v128_t valphav4567 = wasm_v32x4_shuffle(vw4567p0, vw4567p1, 1, 3, 5, 7);

      const v128_t vtltr01 = wasm_f64x2_replace_lane(vtltr0, 1, vtltr1);
      const v128_t vblbr01 = wasm_f64x2_replace_lane(vblbr0, 1, vblbr1);
      const v128_t vtltr23 = wasm_f64x2_replace_lane(vtltr2, 1, vtltr3);
      const v128_t vblbr23 = wasm_f64x2_replace_lane(vblbr2, 1, vblbr3);
      const v128_t vtltr45 = wasm_f64x2_replace_lane(vtltr4, 1, vtltr5);
      const v128_t vblbr45 = wasm_f64x2_replace_lane(vblbr4, 1, vblbr5);
      const v128_t vtltr67 = wasm_f64x2_replace_lane(vtltr6, 1, vtltr7);
      const v128_t vblbr67 = wasm_f64x2_replace_lane(vblbr6, 1, vblbr7);

      const v128_t vldrd01 = wasm_f32x4_sub(vblbr01, vtltr01);
      const v128_t vldrd23 = wasm_f32x4_sub(vblbr23, vtltr23);
      const v128_t vldrd45 = wasm_f32x4_sub(vblbr45, vtltr45);
      const v128_t vldrd67 = wasm_f32x4_sub(vblbr67, vtltr67);

      const v128_t vld0123 = wasm_v32x4_shuffle(vldrd01, vldrd23, 0, 2, 4, 6);
      const v128_t vrd0123 = wasm_v32x4_shuffle(vldrd01, vldrd23, 1, 3, 5, 7);
      const v128_t vld4567 = wasm_v32x4_shuffle(vldrd45, vldrd67, 0, 2, 4, 6);
      const v128_t vrd4567 = wasm_v32x4_shuffle(vldrd45, vldrd67, 1, 3, 5, 7);

      const v128_t vtl0123 = wasm_v32x4_shuffle(vtltr01, vtltr23, 0, 2, 4, 6);
      const v128_t vtr0123 = wasm_v32x4_shuffle(vtltr01, vtltr23, 1, 3, 5, 7);
      const v128_t vtl4567 = wasm_v32x4_shuffle(vtltr45, vtltr67, 0, 2, 4, 6);
      const v128_t vtr4567 = wasm_v32x4_shuffle(vtltr45, vtltr67, 1, 3, 5, 7);

      const v128_t vl0123 = wasm_f32x4_add(vtl0123, wasm_f32x4_mul(vld0123, valphav0123));
      const v128_t vr0123 = wasm_f32x4_add(vtr0123, wasm_f32x4_mul(vrd0123, valphav0123));
      const v128_t vl4567 = wasm_f32x4_add(vtl4567, wasm_f32x4_mul(vld4567, valphav4567));
      const v128_t vr4567 = wasm_f32x4_add(vtr4567, wasm_f32x4_mul(vrd4567, valphav4567));

      const v128_t vd0123 = wasm_f32x4_sub(vr0123, vl0123);
      const v128_t vd4567 = wasm_f32x4_sub(vr4567, vl4567);

      const v128_t vo0123 = wasm_f32x4_add(vl0123, wasm_f32x4_mul(vd0123, valphah0123));
      const v128_t vo4567 = wasm_f32x4_add(vl4567, wasm_f32x4_mul(vd4567, valphah4567));

      wasm_v128_store(output + 0, vo0123);
      wasm_v128_store(output + 4, vo4567);
      output += 8;
    }

    for (; p >= 4; p -= 4) {
      const float* itl0 = (const float*) ((uintptr_t) i[0] + input_offset);
      const float* ibl0 = (const float*) ((uintptr_t) i[1] + input_offset);
      const float* itl1 = (const float*) ((uintptr_t) i[2] + input_offset);
      const float* ibl1 = (const float*) ((uintptr_t) i[3] + input_offset);
      const float* itl2 = (const float*) ((uintptr_t) i[4] + input_offset);
      const float* ibl2 = (const float*) ((uintptr_t) i[5] + input_offset);
      const float* itl3 = (const float*) ((uintptr_t) i[6] + input_offset);
      const float* ibl3 = (const float*) ((uintptr_t) i[7] + input_offset);
      i += 8;

      const v128_t vw0 = wasm_v128_load(w);
      const v128_t vw1 = wasm_v128_load(w + 4);
      w += 8;

      const v128_t vtltr0 = wasm_v128_load64_splat(itl0);
      const v128_t vblbr0 = wasm_v128_load64_splat(ibl0);
      const double vtltr1 = *((const double*) itl1);
      const double vblbr1 = *((const double*) ibl1);
      const v128_t vtltr2 = wasm_v128_load64_splat(itl2);
      const v128_t vblbr2 = wasm_v128_load64_splat(ibl2);
      const double vtltr3 = *((const double*) itl3);
      const double vblbr3 = *((const double*) ibl3);

      const v128_t valphah = wasm_v32x4_shuffle(vw0, vw1, 0, 2, 4, 6);
      const v128_t valphav = wasm_v32x4_shuffle(vw0, vw1, 1, 3, 5, 7);

      const v128_t vtltr01 = wasm_f64x2_replace_lane(vtltr0, 1, vtltr1);
      const v128_t vblbr01 = wasm_f64x2_replace_lane(vblbr0, 1, vblbr1);
      const v128_t vtltr23 = wasm_f64x2_replace_lane(vtltr2, 1, vtltr3);
      const v128_t vblbr23 = wasm_f64x2_replace_lane(vblbr2, 1, vblbr3);

      const v128_t vldrd01 = wasm_f32x4_sub(vblbr01, vtltr01);
      const v128_t vldrd23 = wasm_f32x4_sub(vblbr23, vtltr23);

      const v128_t vld = wasm_v32x4_shuffle(vldrd01, vldrd23, 0, 2, 4, 6);
      const v128_t vrd = wasm_v32x4_shuffle(vldrd01, vldrd23, 1, 3, 5, 7);

      const v128_t vtl = wasm_v32x4_shuffle(vtltr01, vtltr23, 0, 2, 4, 6);
      const v128_t vtr = wasm_v32x4_shuffle(vtltr01, vtltr23, 1, 3, 5, 7);

      const v128_t vl = wasm_f32x4_add(vtl, wasm_f32x4_mul(vld, valphav));
      const v128_t vr = wasm_f32x4_add(vtr, wasm_f32x4_mul(vrd, valphav));

      const v128_t vd = wasm_f32x4_sub(vr, vl);
      const v128_t vo = wasm_f32x4_add(vl, wasm_f32x4_mul(vd, valphah));

      wasm_v128_store(output, vo);
      output += 4;
    }

    if XNN_UNLIKELY(p != 0) {
      if (p & 2) {
        const v128_t vw = wasm_v128_load(w);
        w += 4;

        const v128_t valphah = wasm_v32x4_shuffle(vw, vw, 0, 2, 0, 2);
        const v128_t valphav = wasm_v32x4_shuffle(vw, vw, 1, 3, 1, 3);

        const float* itl0 = (const float*) ((uintptr_t) i[0] + input_offset);
        const float* ibl0 = (const float*) ((uintptr_t) i[1] + input_offset);
        const float* itl1 = (const float*) ((uintptr_t) i[2] + input_offset);
        const float* ibl1 = (const float*) ((uintptr_t) i[3] + input_offset);
        i += 4;

        const v128_t vtltr = wasm_f64x2_replace_lane(wasm_v128_load64_splat(itl0), 1, *((const double*) itl1));
        const v128_t vblbr = wasm_f64x2_replace_lane(wasm_v128_load64_splat(ibl0), 1, *((const double*) ibl1));

        const v128_t vldrd = wasm_f32x4_sub(vblbr, vtltr);
        const v128_t vld = wasm_v32x4_shuffle(vldrd, vldrd, 0, 2, 0, 2);
        const v128_t vrd = wasm_v32x4_shuffle(vldrd, vldrd, 1, 3, 1, 3);

        const v128_t vtl = wasm_v32x4_shuffle(vtltr, vtltr, 0, 2, 0, 2);
        const v128_t vtr = wasm_v32x4_shuffle(vtltr, vtltr, 1, 3, 1, 3);

        const v128_t vl = wasm_f32x4_add(vtl, wasm_f32x4_mul(vld, valphav));
        const v128_t vr = wasm_f32x4_add(vtr, wasm_f32x4_mul(vrd, valphav));

        const v128_t vd = wasm_f32x4_sub(vr, vl);
        const v128_t vo = wasm_f32x4_add(vl, wasm_f32x4_mul(vd, valphah));

        *((double*) output) = wasm_f64x2_extract_lane(vo, 0);
        output += 2;
      }

      if (p & 1) {
        // We are computing the following formula:
        //   result = (1 - alpha_h) * (1 - alpha_v) * top_left +
        //                 alpha_h  * (1 - alpha_v) * top_right +
        //            (1 - alpha_h) *      alpha_v  * bottom_left +
        //                 alpha_h  *      alpha_v  * bottom_right.
        //
        // Rearranging gives
        //   result =    left + alpha_h * (right        - left),
        // where
        //   left =  top_left + alpha_v * (bottom_left  - top_left),
        //  right = top_right + alpha_v * (bottom_right - top_right).

        const float alphah = *w;
        const v128_t valphav = wasm_v128_load32_splat(w + 1);
        w += 2;

        const float* itl = (const float*) ((uintptr_t) i[0] + input_offset);
        const float* ibl = (const float*) ((uintptr_t) i[1] + input_offset);
        i += 2;

        const v128_t vtltr = wasm_v128_load64_splat(itl);
        const v128_t vblbr = wasm_v128_load64_splat(ibl);

        // Compute at once
        //    left_diff = bottom_left  - top_left
        //   right_diff = bottom_right - top_right
        const v128_t vldrd = wasm_f32x4_sub(vblbr, vtltr);
        const v128_t vlr = wasm_f32x4_add(vtltr, wasm_f32x4_mul(vldrd, valphav));

        // Extract them and compute the result.
        const float l = wasm_f32x4_extract_lane(vlr, 0);
        const float r = wasm_f32x4_extract_lane(vlr, 1);

        *output++ = l + alphah * (r - l);
      }
    }

    input_offset += input_increment;
  } while (--channels != 0);
}