[haraka-avx512.git] / aes.c


#include "aes.h"


// The lookup-tables are marked const so they can be placed in read-only storage instead of RAM
// The numbers below can be computed dynamically trading ROM for RAM - 
// This can be useful in (embedded) bootloader applications, where ROM is often limited.
static const uint8_t sbox[256] = {
  //0     1    2      3     4    5     6     7      8    9     A      B    C     D     E     F
  0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
  0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
  0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
  0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
  0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
  0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
  0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
  0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
  0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
  0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
  0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
  0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
  0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
  0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
  0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
  0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16 };

#define getSBoxValue(num) (sbox[(num)])


extern __m512i aes2(__m512i s,__m512i key)
{
	alignas(64) uint8_t state[4][4][4];
	_mm512_store_epi32(state[0][0],s);
	alignas(64) uint8_t a[4][4][4];

	const uint8_t U8_INV_FISRT_BIT_MASK[64] = {[0 ... 63] = 0xFE };//mask2
	const uint32_t COEFF[16] = {[0 ... 15] = 0x1b };//mul
	const uint8_t U8_FISRT_BIT_MASK[64] = {[0 ... 63] = 0x01 };//mask
	const uint32_t U32_FISRT_BYTE_MASK[16] = {[0 ... 15] = 0x000000FF };//mask3
	const uint32_t left1[16] = {0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08};//left
	const uint32_t right1[16] = {0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18};//right
	__m512i mask2 = _mm512_load_epi32 (U8_INV_FISRT_BIT_MASK);
	__m512i mul = _mm512_load_epi32 (COEFF);
	__m512i mask = _mm512_load_epi32 (U8_FISRT_BIT_MASK);
	__m512i mask3 = _mm512_load_epi32 (U32_FISRT_BYTE_MASK);
	__m512i left = _mm512_load_epi32 (left1);
	__m512i right = _mm512_load_epi32 (right1);


  	for (uint8_t t = 0; t < 4; ++t)
  		for (uint8_t i = 0; i < 4; ++i)
   	 		for (uint8_t j = 0; j < 4; ++j)
     				 a[t][i][j] = getSBoxValue(state[t][j][i]); // substitute and swap columbs and rows


	__m512i in1 = _mm512_load_epi32 (a[0][0]);
	__m512i mix_left = _mm512_sllv_epi32(in1,left);
	__m512i mix_right = _mm512_srlv_epi32(in1,right);
	__m512i rotated = _mm512_or_si512(mix_right,mix_left);
	_mm512_store_epi32(a[0][0],rotated);


  	for (uint8_t t = 0; t < 4; ++t)
		for(int x = 0; x < 4;x++)
			for(int y = 0; y < 4;y++)
				state[t][y][x] = a[t][x][y]; // unswap colimbs and rows


	__m512i in = _mm512_load_epi32 (state[0][0]);
	__m512i in_shiftR_8 = _mm512_srli_epi32(in,8);
	__m512i xor2 = _mm512_xor_si512(in_shiftR_8,in);
	__m512i in_shiftR_16 = _mm512_srli_epi32(in,16);
	__m512i xor1 = _mm512_xor_si512(in_shiftR_16,xor2);
	__m512i in_shiftR_24 = _mm512_srli_epi32(in,24);
	__m512i xor = _mm512_xor_si512(xor1,in_shiftR_24);

	__m512i xor_byte = _mm512_and_si512(xor,mask3); // make byte
	__m512i xor_byte_coeff = _mm512_mullo_epi32(xor_byte,mask); 

	__m512i in_shiftL_24 = _mm512_slli_epi32(in,24);
	__m512i in_rotL_8 = _mm512_or_si512(in_shiftR_8,in_shiftL_24);
	
	__m512i Tm  = _mm512_xor_si512(in_rotL_8,in);


	__m512i in_shiftR_7 = _mm512_srli_epi32(Tm,7);
	__m512i in_shiftR_7_BYTE = _mm512_and_si512(in_shiftR_7,mask);
	__m512i mul_1b = _mm512_mullo_epi32(in_shiftR_7_BYTE,mul);

	__m512i in_shiftL_1 = _mm512_slli_epi32(Tm,1);
	__m512i in_shiftL_1_BYTE = _mm512_and_si512(in_shiftL_1,mask2);

	__m512i xtime = _mm512_xor_si512(mul_1b,in_shiftL_1_BYTE);
	__m512i tmp = _mm512_xor_si512(xor_byte_coeff,xtime);
	__m512i matrix_multiply = _mm512_xor_si512(tmp,in);
	return _mm512_xor_si512(key,matrix_multiply);
}


extern void aes(_512_state* state,_512_key *k)
{
	__m512i key = _mm512_load_epi32 (k);
	alignas(64) uint8_t a[4][4][4];

	const uint8_t U8_INV_FISRT_BIT_MASK[64] = {[0 ... 63] = 0xFE };//mask2
	const uint32_t COEFF[16] = {[0 ... 15] = 0x1b };//mul
	const uint8_t U8_FISRT_BIT_MASK[64] = {[0 ... 63] = 0x01 };//mask
	const uint32_t U32_FISRT_BYTE_MASK[16] = {[0 ... 15] = 0x000000FF };//mask3
	const uint32_t left1[16] = {0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08};//left
	const uint32_t right1[16] = {0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18};//right
	__m512i mask2 = _mm512_load_epi32 (U8_INV_FISRT_BIT_MASK);
	__m512i mul = _mm512_load_epi32 (COEFF);
	__m512i mask = _mm512_load_epi32 (U8_FISRT_BIT_MASK);
	__m512i mask3 = _mm512_load_epi32 (U32_FISRT_BYTE_MASK);
	__m512i left = _mm512_load_epi32 (left1);
	__m512i right = _mm512_load_epi32 (right1);


  	for (uint8_t t = 0; t < 4; ++t)
  		for (uint8_t i = 0; i < 4; ++i)
   	 		for (uint8_t j = 0; j < 4; ++j)
     				 a[t][i][j] = getSBoxValue((*state)[t][j][i]); // substitute and swap columbs and rows


	__m512i in1 = _mm512_load_epi32 (a[0][0]);
	__m512i mix_left = _mm512_sllv_epi32(in1,left);
	__m512i mix_right = _mm512_srlv_epi32(in1,right);
	__m512i rotated = _mm512_or_si512(mix_right,mix_left);
	_mm512_store_epi32(a[0][0],rotated);


  	for (uint8_t t = 0; t < 4; ++t)
		for(int x = 0; x < 4;x++)
			for(int y = 0; y < 4;y++)
				(*state)[t][y][x] = a[t][x][y]; // unswap colimbs and rows


	__m512i in = _mm512_load_epi32 ((*state)[0][0]);
	__m512i in_shiftR_8 = _mm512_srli_epi32(in,8);
	__m512i xor2 = _mm512_xor_si512(in_shiftR_8,in);
	__m512i in_shiftR_16 = _mm512_srli_epi32(in,16);
	__m512i xor1 = _mm512_xor_si512(in_shiftR_16,xor2);
	__m512i in_shiftR_24 = _mm512_srli_epi32(in,24);
	__m512i xor = _mm512_xor_si512(xor1,in_shiftR_24);

	__m512i xor_byte = _mm512_and_si512(xor,mask3); // make byte
	__m512i xor_byte_coeff = _mm512_mullo_epi32(xor_byte,mask); 

	__m512i in_shiftL_24 = _mm512_slli_epi32(in,24);
	__m512i in_rotL_8 = _mm512_or_si512(in_shiftR_8,in_shiftL_24);
	
	__m512i Tm  = _mm512_xor_si512(in_rotL_8,in);


	__m512i in_shiftR_7 = _mm512_srli_epi32(Tm,7);
	__m512i in_shiftR_7_BYTE = _mm512_and_si512(in_shiftR_7,mask);
	__m512i mul_1b = _mm512_mullo_epi32(in_shiftR_7_BYTE,mul);

	__m512i in_shiftL_1 = _mm512_slli_epi32(Tm,1);
	__m512i in_shiftL_1_BYTE = _mm512_and_si512(in_shiftL_1,mask2);

	__m512i xtime = _mm512_xor_si512(mul_1b,in_shiftL_1_BYTE);
	__m512i tmp = _mm512_xor_si512(xor_byte_coeff,xtime);
	__m512i matrix_multiply = _mm512_xor_si512(tmp,in);
	__m512i ret = _mm512_xor_si512(key,matrix_multiply);
	_mm512_store_epi32((*state)[0][0],ret);
}
Commit	Line	Data
23e3fea5	1
	2	#include "aes.h"
	3
	4
	5
	6
	7
	8
	9
	10	// The lookup-tables are marked const so they can be placed in read-only storage instead of RAM
	11	// The numbers below can be computed dynamically trading ROM for RAM -
	12	// This can be useful in (embedded) bootloader applications, where ROM is often limited.
	13	static const uint8_t sbox[256] = {
	14	//0 1 2 3 4 5 6 7 8 9 A B C D E F
	15	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
	16	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
	17	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
	18	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
	19	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
	20	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
	21	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
	22	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
	23	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
	24	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
	25	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
	26	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
	27	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
	28	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
	29	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
	30	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16 };
	31
	32	#define getSBoxValue(num) (sbox[(num)])
	33
	34
	35	extern __m512i aes2(__m512i s,__m512i key)
	36	{
	37	alignas(64) uint8_t state[4][4][4];
	38	_mm512_store_epi32(state[0][0],s);
	39	alignas(64) uint8_t a[4][4][4];
	40
	41	const uint8_t U8_INV_FISRT_BIT_MASK[64] = {[0 ... 63] = 0xFE };//mask2
afb82921	42	const uint32_t COEFF[16] = {[0 ... 15] = 0x1b };//mul
23e3fea5	43	const uint8_t U8_FISRT_BIT_MASK[64] = {[0 ... 63] = 0x01 };//mask
	44	const uint32_t U32_FISRT_BYTE_MASK[16] = {[0 ... 15] = 0x000000FF };//mask3
	45	const uint32_t left1[16] = {0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08};//left
	46	const uint32_t right1[16] = {0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18};//right
	47	__m512i mask2 = _mm512_load_epi32 (U8_INV_FISRT_BIT_MASK);
	48	__m512i mul = _mm512_load_epi32 (COEFF);
	49	__m512i mask = _mm512_load_epi32 (U8_FISRT_BIT_MASK);
	50	__m512i mask3 = _mm512_load_epi32 (U32_FISRT_BYTE_MASK);
	51	__m512i left = _mm512_load_epi32 (left1);
	52	__m512i right = _mm512_load_epi32 (right1);
	53
	54
	55	for (uint8_t t = 0; t < 4; ++t)
	56	for (uint8_t i = 0; i < 4; ++i)
	57	for (uint8_t j = 0; j < 4; ++j)
	58	a[t][i][j] = getSBoxValue(state[t][j][i]); // substitute and swap columbs and rows
	59
	60
	61	__m512i in1 = _mm512_load_epi32 (a[0][0]);
	62	__m512i mix_left = _mm512_sllv_epi32(in1,left);
	63	__m512i mix_right = _mm512_srlv_epi32(in1,right);
	64	__m512i rotated = _mm512_or_si512(mix_right,mix_left);
	65	_mm512_store_epi32(a[0][0],rotated);
	66
	67
	68	for (uint8_t t = 0; t < 4; ++t)
	69	for(int x = 0; x < 4;x++)
	70	for(int y = 0; y < 4;y++)
	71	state[t][y][x] = a[t][x][y]; // unswap colimbs and rows
	72
	73
	74	__m512i in = _mm512_load_epi32 (state[0][0]);
	75	__m512i in_shiftR_8 = _mm512_srli_epi32(in,8);
	76	__m512i xor2 = _mm512_xor_si512(in_shiftR_8,in);
	77	__m512i in_shiftR_16 = _mm512_srli_epi32(in,16);
	78	__m512i xor1 = _mm512_xor_si512(in_shiftR_16,xor2);
	79	__m512i in_shiftR_24 = _mm512_srli_epi32(in,24);
	80	__m512i xor = _mm512_xor_si512(xor1,in_shiftR_24);
	81
	82	__m512i xor_byte = _mm512_and_si512(xor,mask3); // make byte
	83	__m512i xor_byte_coeff = _mm512_mullo_epi32(xor_byte,mask);
	84
	85	__m512i in_shiftL_24 = _mm512_slli_epi32(in,24);
	86	__m512i in_rotL_8 = _mm512_or_si512(in_shiftR_8,in_shiftL_24);
	87
	88	__m512i Tm = _mm512_xor_si512(in_rotL_8,in);
	89
	90
	91
	92	__m512i in_shiftR_7 = _mm512_srli_epi32(Tm,7);
	93	__m512i in_shiftR_7_BYTE = _mm512_and_si512(in_shiftR_7,mask);
afb82921	94	__m512i mul_1b = _mm512_mullo_epi32(in_shiftR_7_BYTE,mul);
23e3fea5	95
	96	__m512i in_shiftL_1 = _mm512_slli_epi32(Tm,1);
	97	__m512i in_shiftL_1_BYTE = _mm512_and_si512(in_shiftL_1,mask2);
	98
	99	__m512i xtime = _mm512_xor_si512(mul_1b,in_shiftL_1_BYTE);
	100	__m512i tmp = _mm512_xor_si512(xor_byte_coeff,xtime);
	101	__m512i matrix_multiply = _mm512_xor_si512(tmp,in);
	102	return _mm512_xor_si512(key,matrix_multiply);
	103	}
	104
	105
	106
	107	extern void aes(_512_state* state,_512_key *k)
	108	{
	109	__m512i key = _mm512_load_epi32 (k);
	110	alignas(64) uint8_t a[4][4][4];
	111
	112	const uint8_t U8_INV_FISRT_BIT_MASK[64] = {[0 ... 63] = 0xFE };//mask2
afb82921	113	const uint32_t COEFF[16] = {[0 ... 15] = 0x1b };//mul
23e3fea5	114	const uint8_t U8_FISRT_BIT_MASK[64] = {[0 ... 63] = 0x01 };//mask
	115	const uint32_t U32_FISRT_BYTE_MASK[16] = {[0 ... 15] = 0x000000FF };//mask3
	116	const uint32_t left1[16] = {0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08,0x00,0x18,0x10,0x08};//left
	117	const uint32_t right1[16] = {0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18,0x00,0x08,0x10,0x18};//right
	118	__m512i mask2 = _mm512_load_epi32 (U8_INV_FISRT_BIT_MASK);
	119	__m512i mul = _mm512_load_epi32 (COEFF);
	120	__m512i mask = _mm512_load_epi32 (U8_FISRT_BIT_MASK);
	121	__m512i mask3 = _mm512_load_epi32 (U32_FISRT_BYTE_MASK);
	122	__m512i left = _mm512_load_epi32 (left1);
	123	__m512i right = _mm512_load_epi32 (right1);
	124
	125
	126	for (uint8_t t = 0; t < 4; ++t)
	127	for (uint8_t i = 0; i < 4; ++i)
	128	for (uint8_t j = 0; j < 4; ++j)
	129	a[t][i][j] = getSBoxValue((*state)[t][j][i]); // substitute and swap columbs and rows
	130
	131
	132	__m512i in1 = _mm512_load_epi32 (a[0][0]);
	133	__m512i mix_left = _mm512_sllv_epi32(in1,left);
	134	__m512i mix_right = _mm512_srlv_epi32(in1,right);
	135	__m512i rotated = _mm512_or_si512(mix_right,mix_left);
	136	_mm512_store_epi32(a[0][0],rotated);
	137
	138
	139	for (uint8_t t = 0; t < 4; ++t)
	140	for(int x = 0; x < 4;x++)
	141	for(int y = 0; y < 4;y++)
	142	(*state)[t][y][x] = a[t][x][y]; // unswap colimbs and rows
	143
	144
	145	__m512i in = _mm512_load_epi32 ((*state)[0][0]);
	146	__m512i in_shiftR_8 = _mm512_srli_epi32(in,8);
	147	__m512i xor2 = _mm512_xor_si512(in_shiftR_8,in);
	148	__m512i in_shiftR_16 = _mm512_srli_epi32(in,16);
	149	__m512i xor1 = _mm512_xor_si512(in_shiftR_16,xor2);
	150	__m512i in_shiftR_24 = _mm512_srli_epi32(in,24);
	151	__m512i xor = _mm512_xor_si512(xor1,in_shiftR_24);
	152
	153	__m512i xor_byte = _mm512_and_si512(xor,mask3); // make byte
	154	__m512i xor_byte_coeff = _mm512_mullo_epi32(xor_byte,mask);
	155
	156	__m512i in_shiftL_24 = _mm512_slli_epi32(in,24);
	157	__m512i in_rotL_8 = _mm512_or_si512(in_shiftR_8,in_shiftL_24);
	158
	159	__m512i Tm = _mm512_xor_si512(in_rotL_8,in);
	160
	161
	162
	163	__m512i in_shiftR_7 = _mm512_srli_epi32(Tm,7);
	164	__m512i in_shiftR_7_BYTE = _mm512_and_si512(in_shiftR_7,mask);
afb82921	165	__m512i mul_1b = _mm512_mullo_epi32(in_shiftR_7_BYTE,mul);
23e3fea5	166
	167	__m512i in_shiftL_1 = _mm512_slli_epi32(Tm,1);
	168	__m512i in_shiftL_1_BYTE = _mm512_and_si512(in_shiftL_1,mask2);
	169
	170	__m512i xtime = _mm512_xor_si512(mul_1b,in_shiftL_1_BYTE);
	171	__m512i tmp = _mm512_xor_si512(xor_byte_coeff,xtime);
	172	__m512i matrix_multiply = _mm512_xor_si512(tmp,in);
	173	__m512i ret = _mm512_xor_si512(key,matrix_multiply);
	174	_mm512_store_epi32((*state)[0][0],ret);
	175	}
	176
	177