fzorb/REDACTED-rig
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Initial import.

Browse source
This commit is contained in:
XMRig 2017-04-15 09:02:08 +03:00
commit ed320731e9
93 changed files with 21078 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

2
algo/cryptonight/bmi2/CMakeLists.txt Normal file
View file

@ -0,0 +1,2 @@
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mbmi2")
add_library(cryptonight_av3_aesni_bmi2 STATIC ../cryptonight_av3_aesni_bmi2.c)

72
algo/cryptonight/cryptonight.h Normal file
View file

@ -0,0 +1,72 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef __CRYPTONIGHT_H__
#define __CRYPTONIGHT_H__
#include <stddef.h>
#include <stdint.h>
#define MEMORY (1 << 21) /* 2 MiB */
#define MEMORY_M128I (MEMORY >> 4) // 2 MiB / 16 = 128 ki * __m128i
#define ITER (1 << 20)
#define AES_BLOCK_SIZE 16
#define AES_KEY_SIZE 32 /*16*/
#define INIT_SIZE_BLK 8
#define INIT_SIZE_BYTE (INIT_SIZE_BLK * AES_BLOCK_SIZE) // 128
#define INIT_SIZE_M128I (INIT_SIZE_BYTE >> 4) // 8
#pragma pack(push, 1)
union hash_state {
uint8_t b[200];
uint64_t w[25];
};
#pragma pack(pop)
#pragma pack(push, 1)
union cn_slow_hash_state {
union hash_state hs;
struct {
uint8_t k[64];
uint8_t init[INIT_SIZE_BYTE];
};
};
#pragma pack(pop)
struct cryptonight_ctx {
union cn_slow_hash_state state;
uint8_t text[INIT_SIZE_BYTE] __attribute((aligned(16)));
uint64_t a[2] __attribute__((aligned(16)));
uint64_t b[2] __attribute__((aligned(16)));
uint64_t c[2] __attribute__((aligned(16)));
};
extern void (* const extra_hashes[4])(const void *, size_t, char *);
void cryptonight_init(int variant);
void cryptonight_hash(void* output, const void* input, size_t input_len);
int scanhash_cryptonight(int thr_id, uint32_t *hash, uint32_t *restrict pdata, const uint32_t *restrict ptarget, uint32_t max_nonce, unsigned long *restrict hashes_done, const char *memory, struct cryptonight_ctx *persistentctx);
#endif /* __CRYPTONIGHT_H__ */

216
algo/cryptonight/cryptonight_av1_aesni.c Normal file
View file

@ -0,0 +1,216 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
static inline void ExpandAESKey256_sub1(__m128i *tmp1, __m128i *tmp2)
{
__m128i tmp4;
*tmp2 = _mm_shuffle_epi32(*tmp2, 0xFF);
tmp4 = _mm_slli_si128(*tmp1, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
*tmp1 = _mm_xor_si128(*tmp1, *tmp2);
}
static inline void ExpandAESKey256_sub2(__m128i *tmp1, __m128i *tmp3)
{
__m128i tmp2, tmp4;
tmp4 = _mm_aeskeygenassist_si128(*tmp1, 0x00);
tmp2 = _mm_shuffle_epi32(tmp4, 0xAA);
tmp4 = _mm_slli_si128(*tmp3, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
*tmp3 = _mm_xor_si128(*tmp3, tmp2);
}
// Special thanks to Intel for helping me
// with ExpandAESKey256() and its subroutines
static inline void ExpandAESKey256(char *keybuf)
{
__m128i tmp1, tmp2, tmp3, *keys;
keys = (__m128i *)keybuf;
tmp1 = _mm_load_si128((__m128i *)keybuf);
tmp3 = _mm_load_si128((__m128i *)(keybuf+0x10));
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x01);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[2] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[3] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x02);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[4] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[5] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x04);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[6] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[7] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x08);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[8] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[9] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x10);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[10] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[11] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x20);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[12] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[13] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x40);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[14] = tmp1;
}
void cryptonight_av1_aesni(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
keccak((const uint8_t *)input, 76, (uint8_t *) &ctx->state.hs, 200);
uint8_t ExpandedKey[256];
size_t i, j;
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, ctx->state.hs.b, AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
__m128i *longoutput, *expkey, *xmminput;
longoutput = (__m128i *) memory;
expkey = (__m128i *)ExpandedKey;
xmminput = (__m128i *)ctx->text;
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE)
{
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
_mm_store_si128(&(longoutput[(i >> 4)]), xmminput[0]);
_mm_store_si128(&(longoutput[(i >> 4) + 1]), xmminput[1]);
_mm_store_si128(&(longoutput[(i >> 4) + 2]), xmminput[2]);
_mm_store_si128(&(longoutput[(i >> 4) + 3]), xmminput[3]);
_mm_store_si128(&(longoutput[(i >> 4) + 4]), xmminput[4]);
_mm_store_si128(&(longoutput[(i >> 4) + 5]), xmminput[5]);
_mm_store_si128(&(longoutput[(i >> 4) + 6]), xmminput[6]);
_mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]);
}
for (i = 0; i < 2; i++)
{
ctx->a[i] = ((uint64_t *)ctx->state.k)[i] ^ ((uint64_t *)ctx->state.k)[i+4];
ctx->b[i] = ((uint64_t *)ctx->state.k)[i+2] ^ ((uint64_t *)ctx->state.k)[i+6];
}
__m128i a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
__m128i b_x = _mm_load_si128((__m128i *) ctx->b);
uint64_t c[2] __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16)));
for (i = 0; __builtin_expect(i < 0x80000, 1); i++) {
__m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) ctx->a));
_mm_store_si128((__m128i *) c, c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c[0] & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x;
d[0] = d_ptr[0];
d[1] = d_ptr[1];
{
unsigned __int128 res = (unsigned __int128) c[0] * d[0];
d_ptr[0] = ctx->a[0] += res >> 64;
d_ptr[1] = ctx->a[1] += (uint64_t) res;
}
ctx->a[0] ^= d[0];
ctx->a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, &ctx->state.hs.b[32], AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE) {
xmminput[0] = _mm_xor_si128(longoutput[(i >> 4)], xmminput[0]);
xmminput[1] = _mm_xor_si128(longoutput[(i >> 4) + 1], xmminput[1]);
xmminput[2] = _mm_xor_si128(longoutput[(i >> 4) + 2], xmminput[2]);
xmminput[3] = _mm_xor_si128(longoutput[(i >> 4) + 3], xmminput[3]);
xmminput[4] = _mm_xor_si128(longoutput[(i >> 4) + 4], xmminput[4]);
xmminput[5] = _mm_xor_si128(longoutput[(i >> 4) + 5], xmminput[5]);
xmminput[6] = _mm_xor_si128(longoutput[(i >> 4) + 6], xmminput[6]);
xmminput[7] = _mm_xor_si128(longoutput[(i >> 4) + 7], xmminput[7]);
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

239
algo/cryptonight/cryptonight_av1_aesni32.c Normal file
View file

@ -0,0 +1,239 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
static inline uint64_t mul128(uint64_t multiplier, uint64_t multiplicand, uint64_t *product_hi) {
// multiplier = ab = a * 2^32 + b
// multiplicand = cd = c * 2^32 + d
// ab * cd = a * c * 2^64 + (a * d + b * c) * 2^32 + b * d
uint64_t a = multiplier >> 32;
uint64_t b = multiplier & 0xFFFFFFFF;
uint64_t c = multiplicand >> 32;
uint64_t d = multiplicand & 0xFFFFFFFF;
//uint64_t ac = a * c;
uint64_t ad = a * d;
//uint64_t bc = b * c;
uint64_t bd = b * d;
uint64_t adbc = ad + (b * c);
uint64_t adbc_carry = adbc < ad ? 1 : 0;
// multiplier * multiplicand = product_hi * 2^64 + product_lo
uint64_t product_lo = bd + (adbc << 32);
uint64_t product_lo_carry = product_lo < bd ? 1 : 0;
*product_hi = (a * c) + (adbc >> 32) + (adbc_carry << 32) + product_lo_carry;
return product_lo;
}
static inline void ExpandAESKey256_sub1(__m128i *tmp1, __m128i *tmp2)
{
__m128i tmp4;
*tmp2 = _mm_shuffle_epi32(*tmp2, 0xFF);
tmp4 = _mm_slli_si128(*tmp1, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
*tmp1 = _mm_xor_si128(*tmp1, *tmp2);
}
static inline void ExpandAESKey256_sub2(__m128i *tmp1, __m128i *tmp3)
{
__m128i tmp2, tmp4;
tmp4 = _mm_aeskeygenassist_si128(*tmp1, 0x00);
tmp2 = _mm_shuffle_epi32(tmp4, 0xAA);
tmp4 = _mm_slli_si128(*tmp3, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
*tmp3 = _mm_xor_si128(*tmp3, tmp2);
}
// Special thanks to Intel for helping me
// with ExpandAESKey256() and its subroutines
static inline void ExpandAESKey256(char *keybuf)
{
__m128i tmp1, tmp2, tmp3, *keys;
keys = (__m128i *)keybuf;
tmp1 = _mm_load_si128((__m128i *)keybuf);
tmp3 = _mm_load_si128((__m128i *)(keybuf+0x10));
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x01);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[2] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[3] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x02);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[4] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[5] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x04);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[6] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[7] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x08);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[8] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[9] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x10);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[10] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[11] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x20);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[12] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[13] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x40);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[14] = tmp1;
}
void cryptonight_av1_aesni32(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
keccak((const uint8_t *)input, 76, (uint8_t *) &ctx->state.hs, 200);
uint8_t ExpandedKey[256];
size_t i, j;
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, ctx->state.hs.b, AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
__m128i *longoutput, *expkey, *xmminput;
longoutput = (__m128i *) memory;
expkey = (__m128i *)ExpandedKey;
xmminput = (__m128i *)ctx->text;
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE)
{
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
_mm_store_si128(&(longoutput[(i >> 4)]), xmminput[0]);
_mm_store_si128(&(longoutput[(i >> 4) + 1]), xmminput[1]);
_mm_store_si128(&(longoutput[(i >> 4) + 2]), xmminput[2]);
_mm_store_si128(&(longoutput[(i >> 4) + 3]), xmminput[3]);
_mm_store_si128(&(longoutput[(i >> 4) + 4]), xmminput[4]);
_mm_store_si128(&(longoutput[(i >> 4) + 5]), xmminput[5]);
_mm_store_si128(&(longoutput[(i >> 4) + 6]), xmminput[6]);
_mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]);
}
for (i = 0; i < 2; i++)
{
ctx->a[i] = ((uint64_t *)ctx->state.k)[i] ^ ((uint64_t *)ctx->state.k)[i+4];
ctx->b[i] = ((uint64_t *)ctx->state.k)[i+2] ^ ((uint64_t *)ctx->state.k)[i+6];
}
__m128i a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
__m128i b_x = _mm_load_si128((__m128i *) ctx->b);
uint64_t c[2] __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16)));
uint64_t hi;
for (i = 0; __builtin_expect(i < 0x80000, 1); i++) {
__m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) ctx->a));
_mm_store_si128((__m128i *) c, c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c[0] & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x;
d[0] = d_ptr[0];
d[1] = d_ptr[1];
d_ptr[1] = ctx->a[1] += mul128(c[0], d[0], &hi);
d_ptr[0] = ctx->a[0] += hi;
ctx->a[0] ^= d[0];
ctx->a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, &ctx->state.hs.b[32], AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE) {
xmminput[0] = _mm_xor_si128(longoutput[(i >> 4)], xmminput[0]);
xmminput[1] = _mm_xor_si128(longoutput[(i >> 4) + 1], xmminput[1]);
xmminput[2] = _mm_xor_si128(longoutput[(i >> 4) + 2], xmminput[2]);
xmminput[3] = _mm_xor_si128(longoutput[(i >> 4) + 3], xmminput[3]);
xmminput[4] = _mm_xor_si128(longoutput[(i >> 4) + 4], xmminput[4]);
xmminput[5] = _mm_xor_si128(longoutput[(i >> 4) + 5], xmminput[5]);
xmminput[6] = _mm_xor_si128(longoutput[(i >> 4) + 6], xmminput[6]);
xmminput[7] = _mm_xor_si128(longoutput[(i >> 4) + 7], xmminput[7]);
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

237
algo/cryptonight/cryptonight_av2_aesni_wolf.c Normal file
View file

@ -0,0 +1,237 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
static inline void ExpandAESKey256_sub1(__m128i *tmp1, __m128i *tmp2)
{
__m128i tmp4;
*tmp2 = _mm_shuffle_epi32(*tmp2, 0xFF);
tmp4 = _mm_slli_si128(*tmp1, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
*tmp1 = _mm_xor_si128(*tmp1, *tmp2);
}
static inline void ExpandAESKey256_sub2(__m128i *tmp1, __m128i *tmp3)
{
__m128i tmp2, tmp4;
tmp4 = _mm_aeskeygenassist_si128(*tmp1, 0x00);
tmp2 = _mm_shuffle_epi32(tmp4, 0xAA);
tmp4 = _mm_slli_si128(*tmp3, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
*tmp3 = _mm_xor_si128(*tmp3, tmp2);
}
// Special thanks to Intel for helping me
// with ExpandAESKey256() and its subroutines
static inline void ExpandAESKey256(char *keybuf)
{
__m128i tmp1, tmp2, tmp3, *keys;
keys = (__m128i *)keybuf;
tmp1 = _mm_load_si128((__m128i *)keybuf);
tmp3 = _mm_load_si128((__m128i *)(keybuf+0x10));
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x01);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[2] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[3] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x02);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[4] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[5] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x04);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[6] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[7] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x08);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[8] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[9] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x10);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[10] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[11] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x20);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[12] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[13] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x40);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[14] = tmp1;
}
void cryptonight_av2_aesni_wolf(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
keccak((const uint8_t *) input, 76, (uint8_t *) &ctx->state.hs, 200);
uint8_t ExpandedKey[256];
size_t i, j;
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, ctx->state.hs.b, AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
__m128i *longoutput, *expkey, *xmminput;
longoutput = (__m128i *)memory;
expkey = (__m128i *)ExpandedKey;
xmminput = (__m128i *)ctx->text;
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE)
{
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
_mm_store_si128(&(longoutput[(i >> 4)]), xmminput[0]);
_mm_store_si128(&(longoutput[(i >> 4) + 1]), xmminput[1]);
_mm_store_si128(&(longoutput[(i >> 4) + 2]), xmminput[2]);
_mm_store_si128(&(longoutput[(i >> 4) + 3]), xmminput[3]);
_mm_store_si128(&(longoutput[(i >> 4) + 4]), xmminput[4]);
_mm_store_si128(&(longoutput[(i >> 4) + 5]), xmminput[5]);
_mm_store_si128(&(longoutput[(i >> 4) + 6]), xmminput[6]);
_mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]);
}
for (i = 0; i < 2; i++)
{
ctx->a[i] = ((uint64_t *)ctx->state.k)[i] ^ ((uint64_t *)ctx->state.k)[i+4];
ctx->b[i] = ((uint64_t *)ctx->state.k)[i+2] ^ ((uint64_t *)ctx->state.k)[i+6];
}
__m128i b_x = _mm_load_si128((__m128i *)ctx->b);
uint64_t a[2] __attribute((aligned(16))), b[2] __attribute((aligned(16)));
a[0] = ctx->a[0];
a[1] = ctx->a[1];
for(i = 0; __builtin_expect(i < 0x80000, 1); i++)
{
__m128i c_x = _mm_load_si128((__m128i *)&memory[a[0] & 0x1FFFF0]);
__m128i a_x = _mm_load_si128((__m128i *)a);
uint64_t c[2];
c_x = _mm_aesenc_si128(c_x, a_x);
_mm_store_si128((__m128i *)c, c_x);
__builtin_prefetch(&memory[c[0] & 0x1FFFF0], 0, 1);
b_x = _mm_xor_si128(b_x, c_x);
_mm_store_si128((__m128i *)&memory[a[0] & 0x1FFFF0], b_x);
uint64_t *nextblock = (uint64_t *)&memory[c[0] & 0x1FFFF0];
uint64_t b[2];
b[0] = nextblock[0];
b[1] = nextblock[1];
{
uint64_t hi, lo;
// hi,lo = 64bit x 64bit multiply of c[0] and b[0]
__asm__("mulq %3\n\t"
: "=d" (hi),
"=a" (lo)
: "%a" (c[0]),
"rm" (b[0])
: "cc" );
a[0] += hi;
a[1] += lo;
}
uint64_t *dst = (uint64_t *) &memory[c[0] & 0x1FFFF0];
dst[0] = a[0];
dst[1] = a[1];
a[0] ^= b[0];
a[1] ^= b[1];
b_x = c_x;
__builtin_prefetch(&memory[a[0] & 0x1FFFF0], 0, 3);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, &ctx->state.hs.b[32], AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
//for (i = 0; likely(i < MEMORY); i += INIT_SIZE_BYTE)
// aesni_parallel_xor(&ctx->text, ExpandedKey, &ctx->long_state[i]);
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE)
{
xmminput[0] = _mm_xor_si128(longoutput[(i >> 4)], xmminput[0]);
xmminput[1] = _mm_xor_si128(longoutput[(i >> 4) + 1], xmminput[1]);
xmminput[2] = _mm_xor_si128(longoutput[(i >> 4) + 2], xmminput[2]);
xmminput[3] = _mm_xor_si128(longoutput[(i >> 4) + 3], xmminput[3]);
xmminput[4] = _mm_xor_si128(longoutput[(i >> 4) + 4], xmminput[4]);
xmminput[5] = _mm_xor_si128(longoutput[(i >> 4) + 5], xmminput[5]);
xmminput[6] = _mm_xor_si128(longoutput[(i >> 4) + 6], xmminput[6]);
xmminput[7] = _mm_xor_si128(longoutput[(i >> 4) + 7], xmminput[7]);
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

214
algo/cryptonight/cryptonight_av3_aesni_bmi2.c Normal file
View file

@ -0,0 +1,214 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
static inline void ExpandAESKey256_sub1(__m128i *tmp1, __m128i *tmp2)
{
__m128i tmp4;
*tmp2 = _mm_shuffle_epi32(*tmp2, 0xFF);
tmp4 = _mm_slli_si128(*tmp1, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
*tmp1 = _mm_xor_si128(*tmp1, *tmp2);
}
static inline void ExpandAESKey256_sub2(__m128i *tmp1, __m128i *tmp3)
{
__m128i tmp2, tmp4;
tmp4 = _mm_aeskeygenassist_si128(*tmp1, 0x00);
tmp2 = _mm_shuffle_epi32(tmp4, 0xAA);
tmp4 = _mm_slli_si128(*tmp3, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
*tmp3 = _mm_xor_si128(*tmp3, tmp2);
}
// Special thanks to Intel for helping me
// with ExpandAESKey256() and its subroutines
static inline void ExpandAESKey256(char *keybuf)
{
__m128i tmp1, tmp2, tmp3, *keys;
keys = (__m128i *)keybuf;
tmp1 = _mm_load_si128((__m128i *)keybuf);
tmp3 = _mm_load_si128((__m128i *)(keybuf+0x10));
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x01);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[2] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[3] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x02);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[4] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[5] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x04);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[6] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[7] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x08);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[8] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[9] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x10);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[10] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[11] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x20);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[12] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[13] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x40);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[14] = tmp1;
}
void cryptonight_av3_aesni_bmi2(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
keccak((const uint8_t *) input, 76, (uint8_t *) &ctx->state.hs, 200);
uint8_t ExpandedKey[256];
size_t i, j;
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, ctx->state.hs.b, AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
__m128i *longoutput, *expkey, *xmminput;
longoutput = (__m128i *) memory;
expkey = (__m128i *)ExpandedKey;
xmminput = (__m128i *)ctx->text;
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE)
{
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
_mm_store_si128(&(longoutput[(i >> 4)]), xmminput[0]);
_mm_store_si128(&(longoutput[(i >> 4) + 1]), xmminput[1]);
_mm_store_si128(&(longoutput[(i >> 4) + 2]), xmminput[2]);
_mm_store_si128(&(longoutput[(i >> 4) + 3]), xmminput[3]);
_mm_store_si128(&(longoutput[(i >> 4) + 4]), xmminput[4]);
_mm_store_si128(&(longoutput[(i >> 4) + 5]), xmminput[5]);
_mm_store_si128(&(longoutput[(i >> 4) + 6]), xmminput[6]);
_mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]);
}
for (i = 0; i < 2; i++)
{
ctx->a[i] = ((uint64_t *)ctx->state.k)[i] ^ ((uint64_t *)ctx->state.k)[i+4];
ctx->b[i] = ((uint64_t *)ctx->state.k)[i+2] ^ ((uint64_t *)ctx->state.k)[i+6];
}
__m128i a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
__m128i b_x = _mm_load_si128((__m128i *) ctx->b);
uint64_t c[2] __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16)));
uint64_t hi;
for (i = 0; __builtin_expect(i < 0x80000, 1); i++) {
__m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) ctx->a));
_mm_store_si128((__m128i *) c, c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c[0] & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x;
d[0] = d_ptr[0];
d[1] = d_ptr[1];
d_ptr[1] = ctx->a[1] += _mulx_u64(c[0], d[0], &hi);
d_ptr[0] = ctx->a[0] += hi;
ctx->a[0] ^= d[0];
ctx->a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, &ctx->state.hs.b[32], AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
for (i = 0; __builtin_expect(i < MEMORY, 1); i += INIT_SIZE_BYTE) {
xmminput[0] = _mm_xor_si128(longoutput[(i >> 4)], xmminput[0]);
xmminput[1] = _mm_xor_si128(longoutput[(i >> 4) + 1], xmminput[1]);
xmminput[2] = _mm_xor_si128(longoutput[(i >> 4) + 2], xmminput[2]);
xmminput[3] = _mm_xor_si128(longoutput[(i >> 4) + 3], xmminput[3]);
xmminput[4] = _mm_xor_si128(longoutput[(i >> 4) + 4], xmminput[4]);
xmminput[5] = _mm_xor_si128(longoutput[(i >> 4) + 5], xmminput[5]);
xmminput[6] = _mm_xor_si128(longoutput[(i >> 4) + 6], xmminput[6]);
xmminput[7] = _mm_xor_si128(longoutput[(i >> 4) + 7], xmminput[7]);
for(j = 0; j < 10; j++)
{
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

151
algo/cryptonight/cryptonight_av4_legacy.c Normal file
View file

@ -0,0 +1,151 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "compat.h"
#include "crypto/c_keccak.h"
#include "crypto/aesb.h"
#include "crypto/oaes_lib.h"
static inline uint64_t mul128(uint64_t multiplier, uint64_t multiplicand, uint64_t *product_hi) {
// multiplier = ab = a * 2^32 + b
// multiplicand = cd = c * 2^32 + d
// ab * cd = a * c * 2^64 + (a * d + b * c) * 2^32 + b * d
uint64_t a = multiplier >> 32;
uint64_t b = multiplier & 0xFFFFFFFF;
uint64_t c = multiplicand >> 32;
uint64_t d = multiplicand & 0xFFFFFFFF;
//uint64_t ac = a * c;
uint64_t ad = a * d;
//uint64_t bc = b * c;
uint64_t bd = b * d;
uint64_t adbc = ad + (b * c);
uint64_t adbc_carry = adbc < ad ? 1 : 0;
// multiplier * multiplicand = product_hi * 2^64 + product_lo
uint64_t product_lo = bd + (adbc << 32);
uint64_t product_lo_carry = product_lo < bd ? 1 : 0;
*product_hi = (a * c) + (adbc >> 32) + (adbc_carry << 32) + product_lo_carry;
return product_lo;
}
static inline void mul_sum_xor_dst(const uint8_t* a, uint8_t* c, uint8_t* dst) {
uint64_t hi, lo = mul128(((uint64_t*) a)[0], ((uint64_t*) dst)[0], &hi) + ((uint64_t*) c)[1];
hi += ((uint64_t*) c)[0];
((uint64_t*) c)[0] = ((uint64_t*) dst)[0] ^ hi;
((uint64_t*) c)[1] = ((uint64_t*) dst)[1] ^ lo;
((uint64_t*) dst)[0] = hi;
((uint64_t*) dst)[1] = lo;
}
static inline void xor_blocks(uint8_t* a, const uint8_t* b) {
((uint64_t*) a)[0] ^= ((uint64_t*) b)[0];
((uint64_t*) a)[1] ^= ((uint64_t*) b)[1];
}
static inline void xor_blocks_dst(const uint8_t* a, const uint8_t* b, uint8_t* dst) {
((uint64_t*) dst)[0] = ((uint64_t*) a)[0] ^ ((uint64_t*) b)[0];
((uint64_t*) dst)[1] = ((uint64_t*) a)[1] ^ ((uint64_t*) b)[1];
}
void cryptonight_av4_legacy(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx) {
oaes_ctx *aes_ctx = (oaes_ctx*) oaes_alloc();
size_t i, j;
keccak((const uint8_t *)input, 76, (uint8_t *) &ctx->state.hs, 200);
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
oaes_key_import_data(aes_ctx, ctx->state.hs.b, AES_KEY_SIZE);
for (i = 0; likely(i < MEMORY); i += INIT_SIZE_BYTE) {
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 0], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 1], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 2], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 3], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 4], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 5], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 6], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 7], aes_ctx->key->exp_data);
memcpy((void *) &memory[i], ctx->text, INIT_SIZE_BYTE);
}
xor_blocks_dst(&ctx->state.k[0], &ctx->state.k[32], (uint8_t*) ctx->a);
xor_blocks_dst(&ctx->state.k[16], &ctx->state.k[48], (uint8_t*) ctx->b);
for (i = 0; likely(i < ITER / 4); ++i) {
/* Dependency chain: address -> read value ------+
* written value <-+ hard function (AES or MUL) <+
* next address <-+
*/
/* Iteration 1 */
j = ctx->a[0] & 0x1FFFF0;
aesb_single_round((const uint8_t*) &memory[j], (uint8_t *) ctx->c, (const uint8_t *) ctx->a);
xor_blocks_dst((const uint8_t*) ctx->c, (const uint8_t*) ctx->b, (uint8_t*) &memory[j]);
/* Iteration 2 */
mul_sum_xor_dst((const uint8_t*) ctx->c, (uint8_t*) ctx->a, (uint8_t*) &memory[ctx->c[0] & 0x1FFFF0]);
/* Iteration 3 */
j = ctx->a[0] & 0x1FFFF0;
aesb_single_round(&memory[j], (uint8_t *) ctx->b, (uint8_t *) ctx->a);
xor_blocks_dst((const uint8_t*) ctx->b, (const uint8_t*) ctx->c, (uint8_t*) &memory[j]);
/* Iteration 4 */
mul_sum_xor_dst((const uint8_t*) ctx->b, (uint8_t*) ctx->a, (uint8_t*) &memory[ctx->b[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
oaes_key_import_data(aes_ctx, &ctx->state.hs.b[32], AES_KEY_SIZE);
for (i = 0; likely(i < MEMORY); i += INIT_SIZE_BYTE) {
xor_blocks(&ctx->text[0 * AES_BLOCK_SIZE], &memory[i + 0 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[0 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[1 * AES_BLOCK_SIZE], &memory[i + 1 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[1 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[2 * AES_BLOCK_SIZE], &memory[i + 2 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[2 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[3 * AES_BLOCK_SIZE], &memory[i + 3 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[3 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[4 * AES_BLOCK_SIZE], &memory[i + 4 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[4 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[5 * AES_BLOCK_SIZE], &memory[i + 5 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[5 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[6 * AES_BLOCK_SIZE], &memory[i + 6 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[6 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[7 * AES_BLOCK_SIZE], &memory[i + 7 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[7 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
oaes_free((OAES_CTX **) &aes_ctx);
}

248
algo/cryptonight/cryptonight_av5_aesni_experimental.c Normal file
View file

@ -0,0 +1,248 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
static inline void ExpandAESKey256_sub1(__m128i *tmp1, __m128i *tmp2)
{
__m128i tmp4;
*tmp2 = _mm_shuffle_epi32(*tmp2, 0xFF);
tmp4 = _mm_slli_si128(*tmp1, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp1 = _mm_xor_si128(*tmp1, tmp4);
*tmp1 = _mm_xor_si128(*tmp1, *tmp2);
}
static inline void ExpandAESKey256_sub2(__m128i *tmp1, __m128i *tmp3)
{
__m128i tmp2, tmp4;
tmp4 = _mm_aeskeygenassist_si128(*tmp1, 0x00);
tmp2 = _mm_shuffle_epi32(tmp4, 0xAA);
tmp4 = _mm_slli_si128(*tmp3, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
*tmp3 = _mm_xor_si128(*tmp3, tmp4);
*tmp3 = _mm_xor_si128(*tmp3, tmp2);
}
// Special thanks to Intel for helping me
// with ExpandAESKey256() and its subroutines
static inline void ExpandAESKey256(char *keybuf)
{
__m128i tmp1, tmp2, tmp3, *keys;
keys = (__m128i *)keybuf;
tmp1 = _mm_load_si128((__m128i *)keybuf);
tmp3 = _mm_load_si128((__m128i *)(keybuf+0x10));
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x01);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[2] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[3] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x02);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[4] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[5] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x04);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[6] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[7] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x08);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[8] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[9] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x10);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[10] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[11] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x20);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[12] = tmp1;
ExpandAESKey256_sub2(&tmp1, &tmp3);
keys[13] = tmp3;
tmp2 = _mm_aeskeygenassist_si128(tmp3, 0x40);
ExpandAESKey256_sub1(&tmp1, &tmp2);
keys[14] = tmp1;
}
void cryptonight_av5_aesni_experimental(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
keccak((const uint8_t *)input, 76, (uint8_t *) &ctx->state.hs, 200);
uint8_t ExpandedKey[256];
size_t i, j;
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, ctx->state.hs.b, AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
__m128i *longoutput, *expkey, *xmminput;
longoutput = (__m128i *) memory;
expkey = (__m128i *) ExpandedKey;
xmminput = (__m128i *)ctx->text;
// prefetch expkey, all of xmminput and enough longoutput for 4 loops
_mm_prefetch(xmminput, _MM_HINT_T0 );
_mm_prefetch(xmminput + 4, _MM_HINT_T0 );
for (i = 0; i < 64; i += 16) {
_mm_prefetch(longoutput + i, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 4, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 8, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 12, _MM_HINT_T0);
}
_mm_prefetch(expkey, _MM_HINT_T0);
_mm_prefetch(expkey + 4, _MM_HINT_T0);
_mm_prefetch(expkey + 8, _MM_HINT_T0);
for (i = 0; __builtin_expect(i < MEMORY_M128I, 1); i += INIT_SIZE_M128I) {
__builtin_prefetch(longoutput + i + 64, 1, 0);
__builtin_prefetch(longoutput + i + 68, 1, 0);
for(j = 0; j < 10; j++) {
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
_mm_store_si128(&(longoutput[i ]), xmminput[0]);
_mm_store_si128(&(longoutput[i + 1 ]), xmminput[1]);
_mm_store_si128(&(longoutput[i + 2 ]), xmminput[2]);
_mm_store_si128(&(longoutput[i + 3 ]), xmminput[3]);
_mm_store_si128(&(longoutput[i + 4 ]), xmminput[4]);
_mm_store_si128(&(longoutput[i + 5 ]), xmminput[5]);
_mm_store_si128(&(longoutput[i + 6 ]), xmminput[6]);
_mm_store_si128(&(longoutput[i + 7 ]), xmminput[7]);
}
ctx->a[0] = ((uint64_t *) ctx->state.k)[0] ^ ((uint64_t *) ctx->state.k)[4];
ctx->b[0] = ((uint64_t *) ctx->state.k)[2] ^ ((uint64_t *) ctx->state.k)[6];
ctx->a[1] = ((uint64_t *) ctx->state.k)[1] ^ ((uint64_t *) ctx->state.k)[5];
ctx->b[1] = ((uint64_t *) ctx->state.k)[3] ^ ((uint64_t *) ctx->state.k)[7];
__m128i a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
__m128i b_x = _mm_load_si128((__m128i *) ctx->b);
uint64_t c[2] __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16)));
for (i = 0; __builtin_expect(i < 0x80000, 1); i++) {
__m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) ctx->a));
_mm_store_si128((__m128i *) c, c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c[0] & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x;
d[0] = d_ptr[0];
d[1] = d_ptr[1];
{
unsigned __int128 res = (unsigned __int128) c[0] * d[0];
d_ptr[0] = ctx->a[0] += res >> 64;
d_ptr[1] = ctx->a[1] += (uint64_t) res;
}
ctx->a[0] ^= d[0];
ctx->a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
memcpy(ExpandedKey, &ctx->state.hs.b[32], AES_KEY_SIZE);
ExpandAESKey256(ExpandedKey);
_mm_prefetch(xmminput, _MM_HINT_T0 );
_mm_prefetch(xmminput + 4, _MM_HINT_T0 );
for (i = 0; i < 64; i += 16) {
_mm_prefetch(longoutput + i, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 4, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 8, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 12, _MM_HINT_T0);
}
_mm_prefetch(expkey, _MM_HINT_T0);
_mm_prefetch(expkey + 4, _MM_HINT_T0);
_mm_prefetch(expkey + 8, _MM_HINT_T0);
for (i = 0; __builtin_expect(i < MEMORY_M128I, 1); i += INIT_SIZE_M128I) {
_mm_prefetch(longoutput + i + 64, _MM_HINT_T0);
_mm_prefetch(longoutput + i + 68, _MM_HINT_T0);
xmminput[0] = _mm_xor_si128(longoutput[i ], xmminput[0]);
xmminput[1] = _mm_xor_si128(longoutput[i + 1], xmminput[1]);
xmminput[2] = _mm_xor_si128(longoutput[i + 2], xmminput[2]);
xmminput[3] = _mm_xor_si128(longoutput[i + 3], xmminput[3]);
xmminput[4] = _mm_xor_si128(longoutput[i + 4], xmminput[4]);
xmminput[5] = _mm_xor_si128(longoutput[i + 5], xmminput[5]);
xmminput[6] = _mm_xor_si128(longoutput[i + 6], xmminput[6]);
xmminput[7] = _mm_xor_si128(longoutput[i + 7], xmminput[7]);
for(j = 0; j < 10; j++) {
xmminput[0] = _mm_aesenc_si128(xmminput[0], expkey[j]);
xmminput[1] = _mm_aesenc_si128(xmminput[1], expkey[j]);
xmminput[2] = _mm_aesenc_si128(xmminput[2], expkey[j]);
xmminput[3] = _mm_aesenc_si128(xmminput[3], expkey[j]);
xmminput[4] = _mm_aesenc_si128(xmminput[4], expkey[j]);
xmminput[5] = _mm_aesenc_si128(xmminput[5], expkey[j]);
xmminput[6] = _mm_aesenc_si128(xmminput[6], expkey[j]);
xmminput[7] = _mm_aesenc_si128(xmminput[7], expkey[j]);
}
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

142
algo/cryptonight/cryptonight_common.c Normal file
View file

@ -0,0 +1,142 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <stdlib.h>
#ifndef BUILD_TEST
# include "xmrig.h"
#endif
#include "crypto/c_groestl.h"
#include "crypto/c_blake256.h"
#include "crypto/c_jh.h"
#include "crypto/c_skein.h"
#include "cryptonight.h"
#include "options.h"
#if defined(__x86_64__)
void cryptonight_av1_aesni(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av2_aesni_wolf(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av3_aesni_bmi2(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av5_aesni_experimental(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
#elif defined(__i386__)
void cryptonight_av1_aesni32(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
#endif
void cryptonight_av4_legacy(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void (*cryptonight_hash_ctx)(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx) = NULL;
void cryptonight_init(int variant)
{
switch (variant) {
#if defined(__x86_64__)
case XMR_VARIANT_AESNI:
cryptonight_hash_ctx = cryptonight_av1_aesni;
break;
case XMR_VARIANT_AESNI_WOLF:
cryptonight_hash_ctx = cryptonight_av2_aesni_wolf;
break;
case XMR_VARIANT_AESNI_BMI2:
cryptonight_hash_ctx = cryptonight_av3_aesni_bmi2;
break;
case XMR_VARIANT_EXPERIMENTAL:
cryptonight_hash_ctx = cryptonight_av5_aesni_experimental;
break;
#elif defined(__i386__)
case XMR_VARIANT_AESNI:
cryptonight_hash_ctx = cryptonight_av1_aesni32;
break;
#endif
case XMR_VARIANT_LEGACY:
cryptonight_hash_ctx = cryptonight_av4_legacy;
break;
default:
break;
}
}
static inline void do_blake_hash(const void* input, size_t len, char* output) {
blake256_hash((uint8_t*)output, input, len);
}
static inline void do_groestl_hash(const void* input, size_t len, char* output) {
groestl(input, len * 8, (uint8_t*)output);
}
static inline void do_jh_hash(const void* input, size_t len, char* output) {
jh_hash(32 * 8, input, 8 * len, (uint8_t*)output);
}
static inline void do_skein_hash(const void* input, size_t len, char* output) {
skein_hash(8 * 32, input, 8 * len, (uint8_t*)output);
}
void (* const extra_hashes[4])(const void *, size_t, char *) = {do_blake_hash, do_groestl_hash, do_jh_hash, do_skein_hash};
void cryptonight_hash(void* output, const void* input, size_t len) {
uint8_t *memory __attribute((aligned(16))) = (uint8_t *) malloc(MEMORY);
struct cryptonight_ctx *ctx = (struct cryptonight_ctx*)malloc(sizeof(struct cryptonight_ctx));
cryptonight_hash_ctx(output, input, memory, ctx);
free(memory);
free(ctx);
}
#ifndef BUILD_TEST
int scanhash_cryptonight(int thr_id, uint32_t *hash, uint32_t *restrict pdata, const uint32_t *restrict ptarget, uint32_t max_nonce, unsigned long *restrict hashes_done, const char *restrict memory, struct cryptonight_ctx *persistentctx) {
uint32_t *nonceptr = (uint32_t*) (((char*)pdata) + 39);
uint32_t n = *nonceptr - 1;
const uint32_t first_nonce = n + 1;
do {
*nonceptr = ++n;
cryptonight_hash_ctx(hash, pdata, memory, persistentctx);
if (unlikely(hash[7] < ptarget[7])) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (likely((n <= max_nonce && !work_restart[thr_id].restart)));
*hashes_done = n - first_nonce + 1;
return 0;
}
#endif