-rw-r--r-- 14387 libntruprime-20240825/crypto_decode/1013x2393/avx/decode.c raw
/* auto-generated; do not edit */
/* 20240812 djb: more cryptoint usage */

#include <immintrin.h>
#include "crypto_decode.h"
#include "crypto_int16.h"
#include "crypto_int32.h"
#define int16 crypto_int16
#define int32 crypto_int32

static inline int16 mullo(int16 x,int16 y)
{
  return x*y;
}

static inline int16 mulhi(int16 x,int16 y)
{
  return (x*(int32)y)>>16;
}

static inline __m256i add(__m256i x,__m256i y)
{
  return _mm256_add_epi16(x,y);
}

static inline __m256i sub(__m256i x,__m256i y)
{
  return _mm256_sub_epi16(x,y);
}

static inline __m256i shiftleftconst(__m256i x,int16 y)
{
  return _mm256_slli_epi16(x,y);
}

static inline __m256i signedshiftrightconst(__m256i x,int16 y)
{
  return _mm256_srai_epi16(x,y);
}

static inline __m256i subconst(__m256i x,int16 y)
{
  return sub(x,_mm256_set1_epi16(y));
}

static inline __m256i mulloconst(__m256i x,int16 y)
{
  return _mm256_mullo_epi16(x,_mm256_set1_epi16(y));
}

static inline __m256i mulhiconst(__m256i x,int16 y)
{
  return _mm256_mulhi_epi16(x,_mm256_set1_epi16(y));
}

static inline __m256i ifgesubconst(__m256i x,int16 y)
{
  __m256i y16 = _mm256_set1_epi16(y);
  __m256i top16 = _mm256_set1_epi16(y-1);
  return sub(x,_mm256_cmpgt_epi16(x,top16) & y16);
}

static inline __m256i ifnegaddconst(__m256i x,int16 y)
{
  return add(x,signedshiftrightconst(x,15) & _mm256_set1_epi16(y));
}

void crypto_decode(void *v,const unsigned char *s)
{
  int16 *R0 = v;
  int16 R1[507],R2[254],R3[127],R4[64],R5[32],R6[16],R7[8],R8[4],R9[2],R10[1];
  long long i;
  int16 a0,a1,a2;
  __m256i A0,A1,A2,S0,S1,B0,B1,C0,C1;
  
  s += crypto_decode_STRBYTES;
  a1 = 0;
  a1 += *--s; /* 0...255 */
  a1 = mulhi(a1,4)-mulhi(mullo(a1,4749),276);
  a1 += *--s; /* -138...393 */
  a1 -= 276; /* -414...117 */
  a1 += 276&crypto_int16_negative_mask(a1); /* -138...275 */
  a1 += 276&crypto_int16_negative_mask(a1); /* 0...275 */
  R10[0] = a1;
  
  /* R10 ------> R9: reconstruct mod 1*[14506]+[1243] */
  
  i = 0;
  s -= 2;
  a2 = a0 = R10[0];
  a0 = mulhi(a0,-6226)-mulhi(mullo(a0,-1157),14506); /* -8810...7253 */
  a0 += s[2*i+1]; /* -8810...7508 */
  a0 = mulhi(a0,-6226)-mulhi(mullo(a0,-1157),14506); /* -7967...8089 */
  a0 += s[2*i+0]; /* -7967...8344 */
  a0 += 14506&crypto_int16_negative_mask(a0); /* 0...14505 */
  a1 = (a2<<15)+(s[2*i+1]<<7)+((s[2*i]-a0)>>1);
  a1 = mullo(a1,253);

  /* invalid inputs might need reduction mod 1243 */
  a1 -= 1243;
  a1 += 1243&crypto_int16_negative_mask(a1);

  R9[0] = a0;
  R9[1] = a1;
  s -= 0;
  
  /* R9 ------> R8: reconstruct mod 3*[1927]+[165] */
  
  i = 0;
  s -= 1;
  a2 = a0 = R9[1];
  a0 = mulhi(a0,754)-mulhi(mullo(a0,-8706),1927); /* -964...1152 */
  a0 += s[1*i+0]; /* -964...1407 */
  a0 += 1927&crypto_int16_negative_mask(a0); /* 0...1926 */
  a1 = (a2<<8)+s[i]-a0;
  a1 = mullo(a1,3639);

  /* invalid inputs might need reduction mod 165 */
  a1 -= 165;
  a1 += 165&crypto_int16_negative_mask(a1);

  R8[2] = a0;
  R8[3] = a1;
  s -= 1;
  for (i = 0;i >= 0;--i) {
    a2 = a0 = R9[i];
    a0 = mulhi(a0,754)-mulhi(mullo(a0,-8706),1927); /* -964...1152 */
    a0 += s[1*i+0]; /* -964...1407 */
    a0 += 1927&crypto_int16_negative_mask(a0); /* 0...1926 */
    a1 = (a2<<8)+s[i]-a0;
    a1 = mullo(a1,3639);

    /* invalid inputs might need reduction mod 1927 */
    a1 -= 1927;
    a1 += 1927&crypto_int16_negative_mask(a1);

    R8[2*i] = a0;
    R8[2*i+1] = a1;
  }
  
  /* R8 ------> R7: reconstruct mod 7*[11236]+[962] */
  
  i = 0;
  s -= 2;
  a2 = a0 = R8[3];
  a0 = mulhi(a0,1868)-mulhi(mullo(a0,-1493),11236); /* -5618...6085 */
  a0 += s[2*i+1]; /* -5618...6340 */
  a0 = mulhi(a0,1868)-mulhi(mullo(a0,-1493),11236); /* -5779...5798 */
  a0 += s[2*i+0]; /* -5779...6053 */
  a0 += 11236&crypto_int16_negative_mask(a0); /* 0...11235 */
  a1 = (a2<<14)+(s[2*i+1]<<6)+((s[2*i]-a0)>>2);
  a1 = mullo(a1,-26807);

  /* invalid inputs might need reduction mod 962 */
  a1 -= 962;
  a1 += 962&crypto_int16_negative_mask(a1);

  R7[6] = a0;
  R7[7] = a1;
  s -= 6;
  for (i = 2;i >= 0;--i) {
    a2 = a0 = R8[i];
    a0 = mulhi(a0,1868)-mulhi(mullo(a0,-1493),11236); /* -5618...6085 */
    a0 += s[2*i+1]; /* -5618...6340 */
    a0 = mulhi(a0,1868)-mulhi(mullo(a0,-1493),11236); /* -5779...5798 */
    a0 += s[2*i+0]; /* -5779...6053 */
    a0 += 11236&crypto_int16_negative_mask(a0); /* 0...11235 */
    a1 = (a2<<14)+(s[2*i+1]<<6)+((s[2*i]-a0)>>2);
    a1 = mullo(a1,-26807);

    /* invalid inputs might need reduction mod 11236 */
    a1 -= 11236;
    a1 += 11236&crypto_int16_negative_mask(a1);

    R7[2*i] = a0;
    R7[2*i+1] = a1;
  }
  
  /* R7 ------> R6: reconstruct mod 15*[106]+[2322] */
  
  i = 0;
  s -= 1;
  a2 = a0 = R7[7];
  a0 = mulhi(a0,-40)-mulhi(mullo(a0,-27204),106); /* -63...53 */
  a0 += s[1*i+0]; /* -63...308 */
  a0 = mulhi(a0,28)-mulhi(mullo(a0,-618),106); /* -54...53 */
  a0 += 106&crypto_int16_negative_mask(a0); /* 0...105 */
  a1 = (a2<<7)+((s[i]-a0)>>1);
  a1 = mullo(a1,21021);

  /* invalid inputs might need reduction mod 2322 */
  a1 -= 2322;
  a1 += 2322&crypto_int16_negative_mask(a1);

  R6[14] = a0;
  R6[15] = a1;
  s -= 0;
  for (i = 6;i >= 0;--i) {
    a2 = a0 = R7[i];
    a0 = mulhi(a0,28)-mulhi(mullo(a0,-618),106); /* -53...60 */
    a0 += 106&crypto_int16_negative_mask(a0); /* 0...105 */
    a1 = (a2-a0)>>1;
    a1 = mullo(a1,21021);

    /* invalid inputs might need reduction mod 106 */
    a1 -= 106;
    a1 += 106&crypto_int16_negative_mask(a1);

    R6[2*i] = a0;
    R6[2*i+1] = a1;
  }
  
  /* R6 ------> R5: reconstruct mod 31*[164]+[3624] */
  
  i = 0;
  s -= 1;
  a2 = a0 = R6[15];
  a0 = mulhi(a0,16)-mulhi(mullo(a0,28772),164); /* -82...86 */
  a0 += s[1*i+0]; /* -82...341 */
  a0 = mulhi(a0,-64)-mulhi(mullo(a0,-400),164); /* -83...82 */
  a0 += 164&crypto_int16_negative_mask(a0); /* 0...163 */
  a1 = (a2<<6)+((s[i]-a0)>>2);
  a1 = mullo(a1,-25575);

  /* invalid inputs might need reduction mod 3624 */
  a1 -= 3624;
  a1 += 3624&crypto_int16_negative_mask(a1);

  R5[30] = a0;
  R5[31] = a1;
  s -= 15;
  for (i = 14;i >= 0;--i) {
    a2 = a0 = R6[i];
    a0 = mulhi(a0,16)-mulhi(mullo(a0,28772),164); /* -82...86 */
    a0 += s[1*i+0]; /* -82...341 */
    a0 = mulhi(a0,-64)-mulhi(mullo(a0,-400),164); /* -83...82 */
    a0 += 164&crypto_int16_negative_mask(a0); /* 0...163 */
    a1 = (a2<<6)+((s[i]-a0)>>2);
    a1 = mullo(a1,-25575);

    /* invalid inputs might need reduction mod 164 */
    a1 -= 164;
    a1 += 164&crypto_int16_negative_mask(a1);

    R5[2*i] = a0;
    R5[2*i+1] = a1;
  }
  
  /* R5 ------> R4: reconstruct mod 63*[3278]+[283] */
  
  i = 0;
  s -= 1;
  a2 = a0 = R5[31];
  a0 = mulhi(a0,412)-mulhi(mullo(a0,-5118),3278); /* -1639...1742 */
  a0 += s[1*i+0]; /* -1639...1997 */
  a0 += 3278&crypto_int16_negative_mask(a0); /* 0...3277 */
  a1 = (a2<<7)+((s[i]-a0)>>1);
  a1 = mullo(a1,-19113);

  /* invalid inputs might need reduction mod 283 */
  a1 -= 283;
  a1 += 283&crypto_int16_negative_mask(a1);

  R4[62] = a0;
  R4[63] = a1;
  s -= 62;
  i = 15;
  for (;;) {
    A2 = A0 = _mm256_loadu_si256((__m256i *) &R5[i]);
    S0 = _mm256_loadu_si256((__m256i *) (s+2*i));
    S1 = _mm256_srli_epi16(S0,8);
    S0 &= _mm256_set1_epi16(255);
    A0 = sub(mulhiconst(A0,412),mulhiconst(mulloconst(A0,-5118),3278)); /* -1639...1742 */
    A0 = add(A0,S1); /* -1639...1997 */
    A0 = sub(mulhiconst(A0,412),mulhiconst(mulloconst(A0,-5118),3278)); /* -1650...1651 */
    A0 = add(A0,S0); /* -1650...1906 */
    A0 = ifnegaddconst(A0,3278); /* 0...3277 */
    A1 = add(add(shiftleftconst(A2,15),shiftleftconst(S1,7)),signedshiftrightconst(sub(S0,A0),1));
    A1 = mulloconst(A1,-19113);

    /* invalid inputs might need reduction mod 3278 */
    A1 = ifgesubconst(A1,3278);

    /* A0: r0r2r4r6r8r10r12r14 r16r18r20r22r24r26r28r30 */
    /* A1: r1r3r5r7r9r11r13r15 r17r19r21r23r25r27r29r31 */
    B0 = _mm256_unpacklo_epi16(A0,A1);
    B1 = _mm256_unpackhi_epi16(A0,A1);
    /* B0: r0r1r2r3r4r5r6r7 r16r17r18r19r20r21r22r23 */
    /* B1: r8r9r10r11r12r13r14r15 r24r25r26r27r28r29r30r31 */
    C0 = _mm256_permute2x128_si256(B0,B1,0x20);
    C1 = _mm256_permute2x128_si256(B0,B1,0x31);
    /* C0: r0r1r2r3r4r5r6r7 r8r9r10r11r12r13r14r15 */
    /* C1: r16r17r18r19r20r21r22r23 r24r25r26r27r28r29r30r31 */
    _mm256_storeu_si256((__m256i *) (&R4[2*i]),C0);
    _mm256_storeu_si256((__m256i *) (16+&R4[2*i]),C1);
    if (!i) break;
    i = -16-((~15)&-i);
  }
  
  /* R4 ------> R3: reconstruct mod 126*[916]+[283] */
  
  R3[126] = R4[63];
  s -= 63;
  i = 47;
  for (;;) {
    A2 = A0 = _mm256_loadu_si256((__m256i *) &R4[i]);
    S0 = _mm256_cvtepu8_epi16(_mm_loadu_si128((__m128i *) (s+i)));
    A0 = sub(mulhiconst(A0,-240),mulhiconst(mulloconst(A0,-18316),916)); /* -518...458 */
    A0 = add(A0,S0); /* -518...713 */
    A0 = ifnegaddconst(A0,916); /* 0...915 */
    A1 = add(shiftleftconst(A2,6),signedshiftrightconst(sub(S0,A0),2));
    A1 = mulloconst(A1,-17171);

    /* invalid inputs might need reduction mod 916 */
    A1 = ifgesubconst(A1,916);

    /* A0: r0r2r4r6r8r10r12r14 r16r18r20r22r24r26r28r30 */
    /* A1: r1r3r5r7r9r11r13r15 r17r19r21r23r25r27r29r31 */
    B0 = _mm256_unpacklo_epi16(A0,A1);
    B1 = _mm256_unpackhi_epi16(A0,A1);
    /* B0: r0r1r2r3r4r5r6r7 r16r17r18r19r20r21r22r23 */
    /* B1: r8r9r10r11r12r13r14r15 r24r25r26r27r28r29r30r31 */
    C0 = _mm256_permute2x128_si256(B0,B1,0x20);
    C1 = _mm256_permute2x128_si256(B0,B1,0x31);
    /* C0: r0r1r2r3r4r5r6r7 r8r9r10r11r12r13r14r15 */
    /* C1: r16r17r18r19r20r21r22r23 r24r25r26r27r28r29r30r31 */
    _mm256_storeu_si256((__m256i *) (&R3[2*i]),C0);
    _mm256_storeu_si256((__m256i *) (16+&R3[2*i]),C1);
    if (!i) break;
    i = -16-((~15)&-i);
  }
  
  /* R3 ------> R2: reconstruct mod 253*[7744]+[2393] */
  
  i = 0;
  s -= 2;
  a2 = a0 = R3[126];
  a0 = mulhi(a0,3712)-mulhi(mullo(a0,-2166),7744); /* -3872...4800 */
  a0 += s[2*i+1]; /* -3872...5055 */
  a0 = mulhi(a0,3712)-mulhi(mullo(a0,-2166),7744); /* -4092...4158 */
  a0 += s[2*i+0]; /* -4092...4413 */
  a0 += 7744&crypto_int16_negative_mask(a0); /* 0...7743 */
  a1 = (a2<<10)+(s[2*i+1]<<2)+((s[2*i]-a0)>>6);
  a1 = mullo(a1,27081);

  /* invalid inputs might need reduction mod 2393 */
  a1 -= 2393;
  a1 += 2393&crypto_int16_negative_mask(a1);

  R2[252] = a0;
  R2[253] = a1;
  s -= 252;
  i = 110;
  for (;;) {
    A2 = A0 = _mm256_loadu_si256((__m256i *) &R3[i]);
    S0 = _mm256_loadu_si256((__m256i *) (s+2*i));
    S1 = _mm256_srli_epi16(S0,8);
    S0 &= _mm256_set1_epi16(255);
    A0 = sub(mulhiconst(A0,3712),mulhiconst(mulloconst(A0,-2166),7744)); /* -3872...4800 */
    A0 = add(A0,S1); /* -3872...5055 */
    A0 = sub(mulhiconst(A0,3712),mulhiconst(mulloconst(A0,-2166),7744)); /* -4092...4158 */
    A0 = add(A0,S0); /* -4092...4413 */
    A0 = ifnegaddconst(A0,7744); /* 0...7743 */
    A1 = add(add(shiftleftconst(A2,10),shiftleftconst(S1,2)),signedshiftrightconst(sub(S0,A0),6));
    A1 = mulloconst(A1,27081);

    /* invalid inputs might need reduction mod 7744 */
    A1 = ifgesubconst(A1,7744);

    /* A0: r0r2r4r6r8r10r12r14 r16r18r20r22r24r26r28r30 */
    /* A1: r1r3r5r7r9r11r13r15 r17r19r21r23r25r27r29r31 */
    B0 = _mm256_unpacklo_epi16(A0,A1);
    B1 = _mm256_unpackhi_epi16(A0,A1);
    /* B0: r0r1r2r3r4r5r6r7 r16r17r18r19r20r21r22r23 */
    /* B1: r8r9r10r11r12r13r14r15 r24r25r26r27r28r29r30r31 */
    C0 = _mm256_permute2x128_si256(B0,B1,0x20);
    C1 = _mm256_permute2x128_si256(B0,B1,0x31);
    /* C0: r0r1r2r3r4r5r6r7 r8r9r10r11r12r13r14r15 */
    /* C1: r16r17r18r19r20r21r22r23 r24r25r26r27r28r29r30r31 */
    _mm256_storeu_si256((__m256i *) (&R2[2*i]),C0);
    _mm256_storeu_si256((__m256i *) (16+&R2[2*i]),C1);
    if (!i) break;
    i = -16-((~15)&-i);
  }
  
  /* R2 ------> R1: reconstruct mod 506*[88]+[2393] */
  
  R1[506] = R2[253];
  s -= 0;
  i = 237;
  for (;;) {
    A2 = A0 = _mm256_loadu_si256((__m256i *) &R2[i]);
    A0 = sub(mulhiconst(A0,-24),mulhiconst(mulloconst(A0,-745),88)); /* -50...44 */
    A0 = ifnegaddconst(A0,88); /* 0...87 */
    A1 = signedshiftrightconst(sub(A2,A0),3);
    A1 = mulloconst(A1,-29789);

    /* invalid inputs might need reduction mod 88 */
    A1 = ifgesubconst(A1,88);

    /* A0: r0r2r4r6r8r10r12r14 r16r18r20r22r24r26r28r30 */
    /* A1: r1r3r5r7r9r11r13r15 r17r19r21r23r25r27r29r31 */
    B0 = _mm256_unpacklo_epi16(A0,A1);
    B1 = _mm256_unpackhi_epi16(A0,A1);
    /* B0: r0r1r2r3r4r5r6r7 r16r17r18r19r20r21r22r23 */
    /* B1: r8r9r10r11r12r13r14r15 r24r25r26r27r28r29r30r31 */
    C0 = _mm256_permute2x128_si256(B0,B1,0x20);
    C1 = _mm256_permute2x128_si256(B0,B1,0x31);
    /* C0: r0r1r2r3r4r5r6r7 r8r9r10r11r12r13r14r15 */
    /* C1: r16r17r18r19r20r21r22r23 r24r25r26r27r28r29r30r31 */
    _mm256_storeu_si256((__m256i *) (&R1[2*i]),C0);
    _mm256_storeu_si256((__m256i *) (16+&R1[2*i]),C1);
    if (!i) break;
    i = -16-((~15)&-i);
  }
  
  /* R1 ------> R0: reconstruct mod 1013*[2393] */
  
  R0[1012] = 3*R1[506]-3588;
  s -= 1012;
  i = 490;
  for (;;) {
    A2 = A0 = _mm256_loadu_si256((__m256i *) &R1[i]);
    S0 = _mm256_loadu_si256((__m256i *) (s+2*i));
    S1 = _mm256_srli_epi16(S0,8);
    S0 &= _mm256_set1_epi16(255);
    A0 = sub(mulhiconst(A0,-107),mulhiconst(mulloconst(A0,-7011),2393)); /* -1224...1196 */
    A0 = add(A0,S1); /* -1224...1451 */
    A0 = sub(mulhiconst(A0,-107),mulhiconst(mulloconst(A0,-7011),2393)); /* -1199...1198 */
    A0 = add(A0,S0); /* -1199...1453 */
    A0 = ifnegaddconst(A0,2393); /* 0...2392 */
    A1 = add(shiftleftconst(S1,8),sub(S0,A0));
    A1 = mulloconst(A1,-20759);

    /* invalid inputs might need reduction mod 2393 */
    A1 = ifgesubconst(A1,2393);

    A0 = mulloconst(A0,3);
    A1 = mulloconst(A1,3);
    A0 = subconst(A0,3588);
    A1 = subconst(A1,3588);
    /* A0: r0r2r4r6r8r10r12r14 r16r18r20r22r24r26r28r30 */
    /* A1: r1r3r5r7r9r11r13r15 r17r19r21r23r25r27r29r31 */
    B0 = _mm256_unpacklo_epi16(A0,A1);
    B1 = _mm256_unpackhi_epi16(A0,A1);
    /* B0: r0r1r2r3r4r5r6r7 r16r17r18r19r20r21r22r23 */
    /* B1: r8r9r10r11r12r13r14r15 r24r25r26r27r28r29r30r31 */
    C0 = _mm256_permute2x128_si256(B0,B1,0x20);
    C1 = _mm256_permute2x128_si256(B0,B1,0x31);
    /* C0: r0r1r2r3r4r5r6r7 r8r9r10r11r12r13r14r15 */
    /* C1: r16r17r18r19r20r21r22r23 r24r25r26r27r28r29r30r31 */
    _mm256_storeu_si256((__m256i *) (&R0[2*i]),C0);
    _mm256_storeu_si256((__m256i *) (16+&R0[2*i]),C1);
    if (!i) break;
    i = -16-((~15)&-i);
  }
}