gfx/ycbcr/yuv_convert_sse2.cpp
author Gaia Pushbot <release+gaiajson@mozilla.com>
Wed, 11 Sep 2013 09:00:25 -0700
changeset 146638 787feb7772d259c6c65133ef475c0c1307ff4882
parent 60428 d580ec700a1190dd833b059874c4f9b3374ac064
child 252629 91d6e262b662a0b4e47358665e222d3927337af9
permissions -rw-r--r--
Bumping gaia.json for 2 gaia-central revision(s) ======== https://hg.mozilla.org/integration/gaia-central/rev/cdf75f3e9e3a Author: Fabien Cazenave <fabien@cazenave.cc> Desc: Merge pull request #12108 from fabi1cazenave/smsExplicitL10nArguments-bug915058 Bug 915058 - [SMS] use more explicit l10n arguments in dialog.js, r=julienw ======== https://hg.mozilla.org/integration/gaia-central/rev/d3eb0645e8d0 Author: Fabien Cazenave <fabien@cazenave.cc> Desc: Bug 915058 - [SMS] use more explicit l10n arguments in dialog.js, r=julienw

// Copyright (c) 2010 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#include <emmintrin.h>
#include "yuv_row.h"

namespace mozilla {
namespace gfx {

// FilterRows combines two rows of the image using linear interpolation.
// SSE2 version does 16 pixels at a time.
void FilterRows_SSE2(uint8* ybuf, const uint8* y0_ptr, const uint8* y1_ptr,
                     int source_width, int source_y_fraction) {
  __m128i zero = _mm_setzero_si128();
  __m128i y1_fraction = _mm_set1_epi16(source_y_fraction);
  __m128i y0_fraction = _mm_set1_epi16(256 - source_y_fraction);

  const __m128i* y0_ptr128 = reinterpret_cast<const __m128i*>(y0_ptr);
  const __m128i* y1_ptr128 = reinterpret_cast<const __m128i*>(y1_ptr);
  __m128i* dest128 = reinterpret_cast<__m128i*>(ybuf);
  __m128i* end128 = reinterpret_cast<__m128i*>(ybuf + source_width);

  do {
    __m128i y0 = _mm_loadu_si128(y0_ptr128);
    __m128i y1 = _mm_loadu_si128(y1_ptr128);
    __m128i y2 = _mm_unpackhi_epi8(y0, zero);
    __m128i y3 = _mm_unpackhi_epi8(y1, zero);
    y0 = _mm_unpacklo_epi8(y0, zero);
    y1 = _mm_unpacklo_epi8(y1, zero);
    y0 = _mm_mullo_epi16(y0, y0_fraction);
    y1 = _mm_mullo_epi16(y1, y1_fraction);
    y2 = _mm_mullo_epi16(y2, y0_fraction);
    y3 = _mm_mullo_epi16(y3, y1_fraction);
    y0 = _mm_add_epi16(y0, y1);
    y2 = _mm_add_epi16(y2, y3);
    y0 = _mm_srli_epi16(y0, 8);
    y2 = _mm_srli_epi16(y2, 8);
    y0 = _mm_packus_epi16(y0, y2);
    *dest128++ = y0;
    ++y0_ptr128;
    ++y1_ptr128;
  } while (dest128 < end128);
}

}
}