ffdecsa/parallel_128_sse2.h

   1 /* FFdecsa -- fast decsa algorithm
   2  *
   3  * Copyright (C) 2007 Dark Avenger
   4  *               2003-2004  fatih89r
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with this program; if not, write to the Free Software
  18  * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  19  */
  20
  21 #include <emmintrin.h>
  22
  23 #define MEMALIGN_VAL 16
  24
  25 union __u128i {
  26         unsigned int u[4];
  27         __m128i v;
  28 };
  29
  30 static const union __u128i ff0 = {{0x00000000U, 0x00000000U, 0x00000000U, 0x00000000U}};
  31 static const union __u128i ff1 = {{0xffffffffU, 0xffffffffU, 0xffffffffU, 0xffffffffU}};
  32
  33 typedef __m128i group;
  34 #define GROUP_PARALLELISM 128
  35 #define FF0() ff0.v
  36 #define FF1() ff1.v
  37 #define FFAND(a,b) _mm_and_si128((a),(b))
  38 #define FFOR(a,b)  _mm_or_si128((a),(b))
  39 #define FFXOR(a,b) _mm_xor_si128((a),(b))
  40 #define FFNOT(a)   _mm_xor_si128((a),FF1())
  41 #define MALLOC(X)  _mm_malloc(X,16)
  42 #define FREE(X)    _mm_free(X)
  43
  44 /* BATCH */
  45
  46 static const union __u128i ff29 = {{0x29292929U, 0x29292929U, 0x29292929U, 0x29292929U}};
  47 static const union __u128i ff02 = {{0x02020202U, 0x02020202U, 0x02020202U, 0x02020202U}};
  48 static const union __u128i ff04 = {{0x04040404U, 0x04040404U, 0x04040404U, 0x04040404U}};
  49 static const union __u128i ff10 = {{0x10101010U, 0x10101010U, 0x10101010U, 0x10101010U}};
  50 static const union __u128i ff40 = {{0x40404040U, 0x40404040U, 0x40404040U, 0x40404040U}};
  51 static const union __u128i ff80 = {{0x80808080U, 0x80808080U, 0x80808080U, 0x80808080U}};
  52
  53 typedef __m128i batch;
  54 #define BYTES_PER_BATCH 16
  55 #define B_FFN_ALL_29() ff29.v
  56 #define B_FFN_ALL_02() ff02.v
  57 #define B_FFN_ALL_04() ff04.v
  58 #define B_FFN_ALL_10() ff10.v
  59 #define B_FFN_ALL_40() ff40.v
  60 #define B_FFN_ALL_80() ff80.v
  61
  62 #define B_FFAND(a,b) FFAND(a,b)
  63 #define B_FFOR(a,b)  FFOR(a,b)
  64 #define B_FFXOR(a,b) FFXOR(a,b)
  65 #define B_FFSH8L(a,n) _mm_slli_epi64((a),(n))
  66 #define B_FFSH8R(a,n) _mm_srli_epi64((a),(n))
  67
  68 #define M_EMPTY()
  69
  70 #undef BEST_SPAN
  71 #define BEST_SPAN            16
  72
  73 #undef XOR_BEST_BY
  74 inline static void XOR_BEST_BY(unsigned char *d, unsigned char *s1, unsigned char *s2)
  75 {
  76         __m128i vs1 = _mm_load_si128((__m128i*)s1);
  77         __m128i vs2 = _mm_load_si128((__m128i*)s2);
  78         vs1 = _mm_xor_si128(vs1, vs2);
  79         _mm_store_si128((__m128i*)d, vs1);
  80 }
  81
  82 #include "fftable.h"