unsafe private static void compress(Blake2sContext *s, byte *input)
{
uint *m = (uint *)input;
#if FAST_SPAN
if (!BitConverter.IsLittleEndian)
{
var span = new ReadOnlySpan <byte>(input, BlockBytes);
m = (uint *)s->b;
for (int i = 0; i < BlockWords; i++)
{
m[i] = BinaryPrimitives.ReadUInt32LittleEndian(span.Slice(i * WordSize, WordSize));
}
}
#endif
#if USE_INTRINSICS
if (Sse41.IsSupported)
{
mixSse41(s, m);
}
else
#endif
mixScalar(s, m);
}
unsafe private static void mixScalar(Blake2sContext *s, uint *m)
{
uint m00 = m[00];
uint m01 = m[01];
uint m02 = m[02];
uint m03 = m[03];
uint m04 = m[04];
uint m05 = m[05];
uint m06 = m[06];
uint m07 = m[07];
uint m08 = m[08];
uint m09 = m[09];
uint m10 = m[10];
uint m11 = m[11];
uint m12 = m[12];
uint m13 = m[13];
uint m14 = m[14];
uint m15 = m[15];
uint v00 = s->h[0];
uint v01 = s->h[1];
uint v02 = s->h[2];
uint v03 = s->h[3];
uint v04 = s->h[4];
uint v05 = s->h[5];
uint v06 = s->h[6];
uint v07 = s->h[7];
uint v08 = 0x6A09E667u;
uint v09 = 0xBB67AE85u;
uint v10 = 0x3C6EF372u;
uint v11 = 0xA54FF53Au;
uint v12 = 0x510E527Fu;
uint v13 = 0x9B05688Cu;
uint v14 = 0x1F83D9ABu;
uint v15 = 0x5BE0CD19u;
v12 ^= s->t[0];
v13 ^= s->t[1];
v14 ^= s->f[0];
//ROUND 1
v00 += m00;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m02;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m04;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m06;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m05;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m07;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m01;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m03;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m08;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m10;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m12;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m14;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m13;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m15;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m09;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m11;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 2
v00 += m14;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m04;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m09;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m13;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m15;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m06;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m10;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m08;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m01;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m00;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m11;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m05;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m07;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m03;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m12;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m02;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 3
v00 += m11;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m12;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m05;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m15;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m02;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m13;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m08;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m00;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m10;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m03;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m07;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m09;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m01;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m04;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m14;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m06;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 4
v00 += m07;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m03;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m13;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m11;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m12;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m14;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m09;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m01;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m02;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m05;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m04;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m15;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m00;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m08;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m06;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m10;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 5
v00 += m09;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m05;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m02;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m10;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m04;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m15;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m00;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m07;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m14;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m11;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m06;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m03;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m08;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m13;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m01;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m12;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 6
v00 += m02;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m06;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m00;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m08;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m11;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m03;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m12;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m10;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m04;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m07;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m15;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m01;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m14;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m09;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m13;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m05;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 7
v00 += m12;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m01;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m14;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m04;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m13;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m10;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m05;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m15;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m00;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m06;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m09;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m08;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m02;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m11;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m07;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m03;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 8
v00 += m13;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m07;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m12;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m03;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m01;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m09;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m11;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m14;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m05;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m15;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m08;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m02;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m06;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m10;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m00;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m04;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 9
v00 += m06;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m14;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m11;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m00;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m03;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m08;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m15;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m09;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m12;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m13;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m01;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m10;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m04;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m05;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m02;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m07;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
//ROUND 10
v00 += m10;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 16) ^ (v12 << 16);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 12) ^ (v04 << 20);
v01 += m08;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 16) ^ (v13 << 16);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 12) ^ (v05 << 20);
v02 += m07;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 16) ^ (v14 << 16);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 12) ^ (v06 << 20);
v03 += m01;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 16) ^ (v15 << 16);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 12) ^ (v07 << 20);
v02 += m06;
v02 += v06;
v14 ^= v02;
v14 = (v14 >> 8) ^ (v14 << 24);
v10 += v14;
v06 ^= v10;
v06 = (v06 >> 7) ^ (v06 << 25);
v03 += m05;
v03 += v07;
v15 ^= v03;
v15 = (v15 >> 8) ^ (v15 << 24);
v11 += v15;
v07 ^= v11;
v07 = (v07 >> 7) ^ (v07 << 25);
v00 += m02;
v00 += v04;
v12 ^= v00;
v12 = (v12 >> 8) ^ (v12 << 24);
v08 += v12;
v04 ^= v08;
v04 = (v04 >> 7) ^ (v04 << 25);
v01 += m04;
v01 += v05;
v13 ^= v01;
v13 = (v13 >> 8) ^ (v13 << 24);
v09 += v13;
v05 ^= v09;
v05 = (v05 >> 7) ^ (v05 << 25);
v00 += m15;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 16) ^ (v15 << 16);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 12) ^ (v05 << 20);
v01 += m09;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 16) ^ (v12 << 16);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 12) ^ (v06 << 20);
v02 += m03;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 16) ^ (v13 << 16);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 12) ^ (v07 << 20);
v03 += m13;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 16) ^ (v14 << 16);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 12) ^ (v04 << 20);
v02 += m12;
v02 += v07;
v13 ^= v02;
v13 = (v13 >> 8) ^ (v13 << 24);
v08 += v13;
v07 ^= v08;
v07 = (v07 >> 7) ^ (v07 << 25);
v03 += m00;
v03 += v04;
v14 ^= v03;
v14 = (v14 >> 8) ^ (v14 << 24);
v09 += v14;
v04 ^= v09;
v04 = (v04 >> 7) ^ (v04 << 25);
v00 += m11;
v00 += v05;
v15 ^= v00;
v15 = (v15 >> 8) ^ (v15 << 24);
v10 += v15;
v05 ^= v10;
v05 = (v05 >> 7) ^ (v05 << 25);
v01 += m14;
v01 += v06;
v12 ^= v01;
v12 = (v12 >> 8) ^ (v12 << 24);
v11 += v12;
v06 ^= v11;
v06 = (v06 >> 7) ^ (v06 << 25);
s->h[0] ^= v00 ^ v08;
s->h[1] ^= v01 ^ v09;
s->h[2] ^= v02 ^ v10;
s->h[3] ^= v03 ^ v11;
s->h[4] ^= v04 ^ v12;
s->h[5] ^= v05 ^ v13;
s->h[6] ^= v06 ^ v14;
s->h[7] ^= v07 ^ v15;
}
unsafe private static void mixSse41(Blake2sContext *s, uint *m)
{
var row1 = Sse2.LoadVector128(s->h);
var row2 = Sse2.LoadVector128(s->h + 4);
var row3 = v128iv0;
var row4 = v128iv1;
row4 = Sse2.Xor(row4, Sse2.LoadVector128(s->t)); // reads into f[] as well
var m0 = Sse2.LoadVector128(m);
var m1 = Sse2.LoadVector128(m + 4);
var m2 = Sse2.LoadVector128(m + 8);
var m3 = Sse2.LoadVector128(m + 12);
var r16 = v128rm0;
var r8 = v128rm1;
//ROUND 1
#if OLD_INTRINSICS
var b0 = Sse.StaticCast <float, uint>(Sse.Shuffle(Sse.StaticCast <uint, float>(m0), Sse.StaticCast <uint, float>(m1), 0b_10_00_10_00));
#else
var b0 = Sse.Shuffle(m0.AsSingle(), m1.AsSingle(), 0b_10_00_10_00).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
b0 = Sse.StaticCast <float, uint>(Sse.Shuffle(Sse.StaticCast <uint, float>(m0), Sse.StaticCast <uint, float>(m1), 0b_11_01_11_01));
#else
b0 = Sse.Shuffle(m0.AsSingle(), m1.AsSingle(), 0b_11_01_11_01).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <float, uint>(Sse.Shuffle(Sse.StaticCast <uint, float>(m2), Sse.StaticCast <uint, float>(m3), 0b_10_00_10_00));
#else
b0 = Sse.Shuffle(m2.AsSingle(), m3.AsSingle(), 0b_10_00_10_00).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
b0 = Sse.StaticCast <float, uint>(Sse.Shuffle(Sse.StaticCast <uint, float>(m2), Sse.StaticCast <uint, float>(m3), 0b_11_01_11_01));
#else
b0 = Sse.Shuffle(m2.AsSingle(), m3.AsSingle(), 0b_11_01_11_01).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 2
#if OLD_INTRINSICS
var t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(m2), 0b_00_00_11_00));
#else
var t0 = Sse41.Blend(m1.AsUInt16(), m2.AsUInt16(), 0b_00_00_11_00).AsUInt32();
#endif
var t1 = Sse2.ShiftLeftLogical128BitLane(m3, 4);
#if OLD_INTRINSICS
var t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_11_11_00_00));
#else
var t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_11_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_01_00_11);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.Shuffle(m2, 0b_00_00_10_00);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(m3), 0b_11_00_00_00));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_11_11_00_00));
#else
t1 = Sse41.Blend(m1.AsUInt16(), m3.AsUInt16(), 0b_11_00_00_00).AsUInt32();
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_11_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_11_00_01);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
t0 = Sse2.ShiftLeftLogical128BitLane(m1, 4);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m2), Sse.StaticCast <uint, ushort>(t0), 0b_00_11_00_00));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(t1), 0b_11_11_00_00));
#else
t1 = Sse41.Blend(m2.AsUInt16(), t0.AsUInt16(), 0b_00_11_00_00).AsUInt32();
t2 = Sse41.Blend(m0.AsUInt16(), t1.AsUInt16(), 0b_11_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_11_00_01);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackHigh(m0, m1);
t1 = Sse2.ShiftLeftLogical128BitLane(m3, 4);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_11_00));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_11_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_11_00_01);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 3
t0 = Sse2.UnpackHigh(m2, m3);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m3), Sse.StaticCast <uint, ushort>(m1), 0b_00_00_11_00));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_11_11));
#else
t1 = Sse41.Blend(m3.AsUInt16(), m1.AsUInt16(), 0b_00_00_11_00).AsUInt32();
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_11_01_00_10);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackLow(m2, m0);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(m0), 0b_11_11_00_00));
#else
t1 = Sse41.Blend(t0.AsUInt16(), m0.AsUInt16(), 0b_11_11_00_00).AsUInt32();
#endif
t2 = Sse2.ShiftLeftLogical128BitLane(m3, 8);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t2), 0b_11_00_00_00));
#else
b0 = Sse41.Blend(t1.AsUInt16(), t2.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m2), 0b_00_11_11_00));
#else
t0 = Sse41.Blend(m0.AsUInt16(), m2.AsUInt16(), 0b_00_11_11_00).AsUInt32();
#endif
t1 = Sse2.ShiftRightLogical128BitLane(m1, 12);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_00_11));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_00_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_01_00_11_10);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.ShiftLeftLogical128BitLane(m3, 4);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m1), 0b_00_11_00_11));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_11_00_00_00));
#else
t1 = Sse41.Blend(m0.AsUInt16(), m1.AsUInt16(), 0b_00_11_00_11).AsUInt32();
t2 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_00_01_10_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 4
t0 = Sse2.UnpackHigh(m0, m1);
t1 = Sse2.UnpackHigh(t0, m2);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(m3), 0b_00_00_11_00));
#else
t2 = Sse41.Blend(t1.AsUInt16(), m3.AsUInt16(), 0b_00_00_11_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_11_01_00_10);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.ShiftLeftLogical128BitLane(m2, 8);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m3), Sse.StaticCast <uint, ushort>(m0), 0b_00_00_11_00));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_11_00_00_00));
#else
t1 = Sse41.Blend(m3.AsUInt16(), m0.AsUInt16(), 0b_00_00_11_00).AsUInt32();
t2 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_00_01_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m1), 0b_00_00_11_11));
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(m3), 0b_11_00_00_00));
#else
t0 = Sse41.Blend(m0.AsUInt16(), m1.AsUInt16(), 0b_00_00_11_11).AsUInt32();
t1 = Sse41.Blend(t0.AsUInt16(), m3.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t1, 0b_11_00_01_10);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackLow(m0, m2);
t1 = Sse2.UnpackHigh(m1, m2);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(t1), Sse.StaticCast <uint, ulong>(t0)));
#else
b0 = Sse2.UnpackLow(t1.AsUInt64(), t0.AsUInt64()).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 5
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(m1), Sse.StaticCast <uint, ulong>(m2)));
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m0), Sse.StaticCast <uint, ulong>(m2)));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_11_00_11));
#else
t0 = Sse2.UnpackLow(m1.AsUInt64(), m2.AsUInt64()).AsUInt32();
t1 = Sse2.UnpackHigh(m0.AsUInt64(), m2.AsUInt64()).AsUInt32();
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_11_00_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_00_01_11);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m1), Sse.StaticCast <uint, ulong>(m3)));
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(m0), Sse.StaticCast <uint, ulong>(m1)));
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_11_00_11));
#else
t0 = Sse2.UnpackHigh(m1.AsUInt64(), m3.AsUInt64()).AsUInt32();
t1 = Sse2.UnpackLow(m0.AsUInt64(), m1.AsUInt64()).AsUInt32();
b0 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_11_00_11).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m3), Sse.StaticCast <uint, ulong>(m1)));
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m2), Sse.StaticCast <uint, ulong>(m0)));
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_00_11_00_11));
#else
t0 = Sse2.UnpackHigh(m3.AsUInt64(), m1.AsUInt64()).AsUInt32();
t1 = Sse2.UnpackHigh(m2.AsUInt64(), m0.AsUInt64()).AsUInt32();
b0 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_00_11_00_11).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m2), 0b_00_00_00_11));
#else
t0 = Sse41.Blend(m0.AsUInt16(), m2.AsUInt16(), 0b_00_00_00_11).AsUInt32();
#endif
t1 = Sse2.ShiftLeftLogical128BitLane(t0, 8);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(m3), 0b_00_00_11_11));
#else
t2 = Sse41.Blend(t1.AsUInt16(), m3.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_01_10_00_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 6
t0 = Sse2.UnpackHigh(m0, m1);
t1 = Sse2.UnpackLow(m0, m2);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(t0), Sse.StaticCast <uint, ulong>(t1)));
#else
b0 = Sse2.UnpackLow(t0.AsUInt64(), t1.AsUInt64()).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.ShiftRightLogical128BitLane(m2, 4);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m3), 0b_00_00_00_11));
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_00_11_11_00));
#else
t1 = Sse41.Blend(m0.AsUInt16(), m3.AsUInt16(), 0b_00_00_00_11).AsUInt32();
b0 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_00_11_11_00).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(m0), 0b_00_00_11_00));
#else
t0 = Sse41.Blend(m1.AsUInt16(), m0.AsUInt16(), 0b_00_00_11_00).AsUInt32();
#endif
t1 = Sse2.ShiftRightLogical128BitLane(m3, 4);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_11_00_00));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_01_10_11_00);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(m1), Sse.StaticCast <uint, ulong>(m2)));
#else
t0 = Sse2.UnpackLow(m1.AsUInt64(), m2.AsUInt64()).AsUInt32();
#endif
t1 = Sse2.Shuffle(m3, 0b_00_10_00_01);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_11_00_11));
#else
b0 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_11_00_11).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 7
t0 = Sse2.ShiftLeftLogical128BitLane(m1, 12);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m3), 0b_00_11_00_11));
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_11_00_00_00));
#else
t1 = Sse41.Blend(m0.AsUInt16(), m3.AsUInt16(), 0b_00_11_00_11).AsUInt32();
b0 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m3), Sse.StaticCast <uint, ushort>(m2), 0b_00_11_00_00));
#else
t0 = Sse41.Blend(m3.AsUInt16(), m2.AsUInt16(), 0b_00_11_00_00).AsUInt32();
#endif
t1 = Sse2.ShiftRightLogical128BitLane(m1, 4);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_00_11));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_00_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_10_01_11_00);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(m0), Sse.StaticCast <uint, ulong>(m2)));
#else
t0 = Sse2.UnpackLow(m0.AsUInt64(), m2.AsUInt64()).AsUInt32();
#endif
t1 = Sse2.ShiftRightLogical128BitLane(m1, 4);
#if OLD_INTRINSICS
b0 = Sse2.Shuffle(Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_11_00)), 0b_10_11_01_00);
#else
b0 = Sse2.Shuffle(Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_11_00).AsUInt32(), 0b_10_11_01_00);
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackHigh(m1, m2);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m0), Sse.StaticCast <uint, ulong>(t0)));
#else
t1 = Sse2.UnpackHigh(m0.AsUInt64(), t0.AsUInt64()).AsUInt32();
#endif
b0 = Sse2.Shuffle(t1, 0b_11_00_01_10);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 8
t0 = Sse2.UnpackHigh(m0, m1);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(m3), 0b_00_00_11_11));
#else
t1 = Sse41.Blend(t0.AsUInt16(), m3.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t1, 0b_10_00_11_01);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m2), Sse.StaticCast <uint, ushort>(m3), 0b_00_11_00_00));
#else
t0 = Sse41.Blend(m2.AsUInt16(), m3.AsUInt16(), 0b_00_11_00_00).AsUInt32();
#endif
t1 = Sse2.ShiftRightLogical128BitLane(m0, 4);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_00_11));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_00_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_01_00_10_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(m0), Sse.StaticCast <uint, ulong>(m3)));
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(m1), Sse.StaticCast <uint, ulong>(m2)));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_11_11_00));
#else
t0 = Sse2.UnpackHigh(m0.AsUInt64(), m3.AsUInt64()).AsUInt32();
t1 = Sse2.UnpackLow(m1.AsUInt64(), m2.AsUInt64()).AsUInt32();
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_11_11_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_00_10_11_01);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackLow(m0, m1);
t1 = Sse2.UnpackHigh(m1, m2);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(t0), Sse.StaticCast <uint, ulong>(t1)));
#else
b0 = Sse2.UnpackLow(t0.AsUInt64(), t1.AsUInt64()).AsUInt32();
#endif
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 9
t0 = Sse2.UnpackHigh(m1, m3);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ulong, uint>(Sse2.UnpackLow(Sse.StaticCast <uint, ulong>(t0), Sse.StaticCast <uint, ulong>(m0)));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(m2), 0b_11_00_00_00));
b0 = Sse.StaticCast <ushort, uint>(Sse2.ShuffleHigh(Sse.StaticCast <uint, ushort>(t2), 0b_01_00_11_10));
#else
t1 = Sse2.UnpackLow(t0.AsUInt64(), m0.AsUInt64()).AsUInt32();
t2 = Sse41.Blend(t1.AsUInt16(), m2.AsUInt16(), 0b_11_00_00_00).AsUInt32();
b0 = Sse2.ShuffleHigh(t2.AsUInt16(), 0b_01_00_11_10).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.UnpackHigh(m0, m3);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m2), Sse.StaticCast <uint, ushort>(t0), 0b_11_11_00_00));
#else
t1 = Sse41.Blend(m2.AsUInt16(), t0.AsUInt16(), 0b_11_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t1, 0b_00_10_01_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m2), Sse.StaticCast <uint, ushort>(m0), 0b_00_00_11_00));
#else
t0 = Sse41.Blend(m2.AsUInt16(), m0.AsUInt16(), 0b_00_00_11_00).AsUInt32();
#endif
t1 = Sse2.ShiftLeftLogical128BitLane(t0, 4);
#if OLD_INTRINSICS
b0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(m3), 0b_00_00_11_11));
#else
b0 = Sse41.Blend(t1.AsUInt16(), m3.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(m0), 0b_00_11_00_00));
#else
t0 = Sse41.Blend(m1.AsUInt16(), m0.AsUInt16(), 0b_00_11_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t0, 0b_01_00_11_10);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
//ROUND 10
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m0), Sse.StaticCast <uint, ushort>(m2), 0b_00_00_00_11));
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(m2), 0b_00_11_00_00));
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t1), Sse.StaticCast <uint, ushort>(t0), 0b_00_00_11_11));
#else
t0 = Sse41.Blend(m0.AsUInt16(), m2.AsUInt16(), 0b_00_00_00_11).AsUInt32();
t1 = Sse41.Blend(m1.AsUInt16(), m2.AsUInt16(), 0b_00_11_00_00).AsUInt32();
t2 = Sse41.Blend(t1.AsUInt16(), t0.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_01_11_00_10);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
t0 = Sse2.ShiftLeftLogical128BitLane(m0, 4);
#if OLD_INTRINSICS
t1 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m1), Sse.StaticCast <uint, ushort>(t0), 0b_11_00_00_00));
#else
t1 = Sse41.Blend(m1.AsUInt16(), t0.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
b0 = Sse2.Shuffle(t1, 0b_01_10_00_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//DIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_10_01_00_11);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_00_11_10_01);
t0 = Sse2.UnpackHigh(m0, m3);
t1 = Sse2.UnpackLow(m2, m3);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ulong, uint>(Sse2.UnpackHigh(Sse.StaticCast <uint, ulong>(t0), Sse.StaticCast <uint, ulong>(t1)));
#else
t2 = Sse2.UnpackHigh(t0.AsUInt64(), t1.AsUInt64()).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_11_00_10_01);
//G1
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r16));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r16).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 12), Sse2.ShiftLeftLogical(row2, 20));
#if OLD_INTRINSICS
t0 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(m3), Sse.StaticCast <uint, ushort>(m2), 0b_11_00_00_00));
#else
t0 = Sse41.Blend(m3.AsUInt16(), m2.AsUInt16(), 0b_11_00_00_00).AsUInt32();
#endif
t1 = Sse2.UnpackLow(m0, m3);
#if OLD_INTRINSICS
t2 = Sse.StaticCast <ushort, uint>(Sse41.Blend(Sse.StaticCast <uint, ushort>(t0), Sse.StaticCast <uint, ushort>(t1), 0b_00_00_11_11));
#else
t2 = Sse41.Blend(t0.AsUInt16(), t1.AsUInt16(), 0b_00_00_11_11).AsUInt32();
#endif
b0 = Sse2.Shuffle(t2, 0b_00_01_10_11);
//G2
row1 = Sse2.Add(Sse2.Add(row1, b0), row2);
row4 = Sse2.Xor(row4, row1);
#if OLD_INTRINSICS
row4 = Sse.StaticCast <sbyte, uint>(Ssse3.Shuffle(Sse.StaticCast <uint, sbyte>(row4), r8));
#else
row4 = Ssse3.Shuffle(row4.AsSByte(), r8).AsUInt32();
#endif
row3 = Sse2.Add(row3, row4);
row2 = Sse2.Xor(row2, row3);
row2 = Sse2.Xor(Sse2.ShiftRightLogical(row2, 7), Sse2.ShiftLeftLogical(row2, 25));
//UNDIAGONALIZE
row4 = Sse2.Shuffle(row4, 0b_00_11_10_01);
row3 = Sse2.Shuffle(row3, 0b_01_00_11_10);
row2 = Sse2.Shuffle(row2, 0b_10_01_00_11);
row1 = Sse2.Xor(row1, row3);
row2 = Sse2.Xor(row2, row4);
row1 = Sse2.Xor(row1, Sse2.LoadVector128(s->h));
row2 = Sse2.Xor(row2, Sse2.LoadVector128(s->h + 4));
Sse2.Store(s->h, row1);
Sse2.Store(s->h + 4, row2);
}