#define ROL32(x, y)		rotate(x, y ## U)
#define SHR(x, y)		(x >> y)
#define SWAP32(a)    	(as_uint(as_uchar4(a).wzyx))

#define S0(x) (ROL32(x, 25) ^ ROL32(x, 14) ^  SHR(x, 3))
#define S1(x) (ROL32(x, 15) ^ ROL32(x, 13) ^  SHR(x, 10))

#define S2(x) (ROL32(x, 30) ^ ROL32(x, 19) ^ ROL32(x, 10))
#define S3(x) (ROL32(x, 26) ^ ROL32(x, 21) ^ ROL32(x, 7))

#define P(a,b,c,d,e,f,g,h,x,K)                  \
{                                               \
    temp1 = h + S3(e) + F1(e,f,g) + (K + x);      \
    d += temp1; h = temp1 + S2(a) + F0(a,b,c);  \
}

#define F0(y, x, z) bitselect(z, y, z ^ x)
#define F1(x, y, z) bitselect(z, y, x)

#define R0 (W0 = S1(W14) + W9 + S0(W1) + W0)
#define R1 (W1 = S1(W15) + W10 + S0(W2) + W1)
#define R2 (W2 = S1(W0) + W11 + S0(W3) + W2)
#define R3 (W3 = S1(W1) + W12 + S0(W4) + W3)
#define R4 (W4 = S1(W2) + W13 + S0(W5) + W4)
#define R5 (W5 = S1(W3) + W14 + S0(W6) + W5)
#define R6 (W6 = S1(W4) + W15 + S0(W7) + W6)
#define R7 (W7 = S1(W5) + W0 + S0(W8) + W7)
#define R8 (W8 = S1(W6) + W1 + S0(W9) + W8)
#define R9 (W9 = S1(W7) + W2 + S0(W10) + W9)
#define R10 (W10 = S1(W8) + W3 + S0(W11) + W10)
#define R11 (W11 = S1(W9) + W4 + S0(W12) + W11)
#define R12 (W12 = S1(W10) + W5 + S0(W13) + W12)
#define R13 (W13 = S1(W11) + W6 + S0(W14) + W13)
#define R14 (W14 = S1(W12) + W7 + S0(W15) + W14)
#define R15 (W15 = S1(W13) + W8 + S0(W0) + W15)

#define RD14 (S1(W12) + W7 + S0(W15) + W14)
#define RD15 (S1(W13) + W8 + S0(W0) + W15)


inline uint8 sha256_round(uint16 data, uint8 buf)
{
	uint temp1;
	uint8 res;
	uint W0 = (data.s0);
	uint W1 = (data.s1);
	uint W2 = (data.s2);
	uint W3 = (data.s3);
	uint W4 = (data.s4);
	uint W5 = (data.s5);
	uint W6 = (data.s6);
	uint W7 = (data.s7);
	uint W8 = (data.s8);
	uint W9 = (data.s9);
	uint W10 = (data.sA);
	uint W11 = (data.sB);
	uint W12 = (data.sC);
	uint W13 = (data.sD);
	uint W14 = (data.sE);
	uint W15 = (data.sF);

	uint v0 = buf.s0;
	uint v1 = buf.s1;
	uint v2 = buf.s2;
	uint v3 = buf.s3;
	uint v4 = buf.s4;
	uint v5 = buf.s5;
	uint v6 = buf.s6;
	uint v7 = buf.s7;

	P(v0, v1, v2, v3, v4, v5, v6, v7, W0, 0x428A2F98);
	P(v7, v0, v1, v2, v3, v4, v5, v6, W1, 0x71374491);
	P(v6, v7, v0, v1, v2, v3, v4, v5, W2, 0xB5C0FBCF);
	P(v5, v6, v7, v0, v1, v2, v3, v4, W3, 0xE9B5DBA5);
	P(v4, v5, v6, v7, v0, v1, v2, v3, W4, 0x3956C25B);
	P(v3, v4, v5, v6, v7, v0, v1, v2, W5, 0x59F111F1);
	P(v2, v3, v4, v5, v6, v7, v0, v1, W6, 0x923F82A4);
	P(v1, v2, v3, v4, v5, v6, v7, v0, W7, 0xAB1C5ED5);
	P(v0, v1, v2, v3, v4, v5, v6, v7, W8, 0xD807AA98);
	P(v7, v0, v1, v2, v3, v4, v5, v6, W9, 0x12835B01);
	P(v6, v7, v0, v1, v2, v3, v4, v5, W10, 0x243185BE);
	P(v5, v6, v7, v0, v1, v2, v3, v4, W11, 0x550C7DC3);
	P(v4, v5, v6, v7, v0, v1, v2, v3, W12, 0x72BE5D74);
	P(v3, v4, v5, v6, v7, v0, v1, v2, W13, 0x80DEB1FE);
	P(v2, v3, v4, v5, v6, v7, v0, v1, W14, 0x9BDC06A7);
	P(v1, v2, v3, v4, v5, v6, v7, v0, W15, 0xC19BF174);

	P(v0, v1, v2, v3, v4, v5, v6, v7, R0, 0xE49B69C1);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R1, 0xEFBE4786);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R2, 0x0FC19DC6);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R3, 0x240CA1CC);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R4, 0x2DE92C6F);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R5, 0x4A7484AA);
	P(v2, v3, v4, v5, v6, v7, v0, v1, R6, 0x5CB0A9DC);
	P(v1, v2, v3, v4, v5, v6, v7, v0, R7, 0x76F988DA);
	P(v0, v1, v2, v3, v4, v5, v6, v7, R8, 0x983E5152);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R9, 0xA831C66D);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R10, 0xB00327C8);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R11, 0xBF597FC7);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R12, 0xC6E00BF3);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R13, 0xD5A79147);
	P(v2, v3, v4, v5, v6, v7, v0, v1, R14, 0x06CA6351);
	P(v1, v2, v3, v4, v5, v6, v7, v0, R15, 0x14292967);

	P(v0, v1, v2, v3, v4, v5, v6, v7, R0, 0x27B70A85);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R1, 0x2E1B2138);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R2, 0x4D2C6DFC);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R3, 0x53380D13);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R4, 0x650A7354);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R5, 0x766A0ABB);
	P(v2, v3, v4, v5, v6, v7, v0, v1, R6, 0x81C2C92E);
	P(v1, v2, v3, v4, v5, v6, v7, v0, R7, 0x92722C85);
	P(v0, v1, v2, v3, v4, v5, v6, v7, R8, 0xA2BFE8A1);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R9, 0xA81A664B);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R10, 0xC24B8B70);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R11, 0xC76C51A3);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R12, 0xD192E819);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R13, 0xD6990624);
	P(v2, v3, v4, v5, v6, v7, v0, v1, R14, 0xF40E3585);
	P(v1, v2, v3, v4, v5, v6, v7, v0, R15, 0x106AA070);

	P(v0, v1, v2, v3, v4, v5, v6, v7, R0, 0x19A4C116);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R1, 0x1E376C08);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R2, 0x2748774C);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R3, 0x34B0BCB5);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R4, 0x391C0CB3);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R5, 0x4ED8AA4A);
	P(v2, v3, v4, v5, v6, v7, v0, v1, R6, 0x5B9CCA4F);
	P(v1, v2, v3, v4, v5, v6, v7, v0, R7, 0x682E6FF3);
	P(v0, v1, v2, v3, v4, v5, v6, v7, R8, 0x748F82EE);
	P(v7, v0, v1, v2, v3, v4, v5, v6, R9, 0x78A5636F);
	P(v6, v7, v0, v1, v2, v3, v4, v5, R10, 0x84C87814);
	P(v5, v6, v7, v0, v1, v2, v3, v4, R11, 0x8CC70208);
	P(v4, v5, v6, v7, v0, v1, v2, v3, R12, 0x90BEFFFA);
	P(v3, v4, v5, v6, v7, v0, v1, v2, R13, 0xA4506CEB);
	P(v2, v3, v4, v5, v6, v7, v0, v1, RD14, 0xBEF9A3F7);
	P(v1, v2, v3, v4, v5, v6, v7, v0, RD15, 0xC67178F2);

	res.s0 = (v0 + buf.s0);
	res.s1 = (v1 + buf.s1);
	res.s2 = (v2 + buf.s2);
	res.s3 = (v3 + buf.s3);
	res.s4 = (v4 + buf.s4);
	res.s5 = (v5 + buf.s5);
	res.s6 = (v6 + buf.s6);
	res.s7 = (v7 + buf.s7);
	return (res);
}