int scanhash_X(int thr_id, uint32_t *pdata, const uint32_t *ptarget, uint32_t max_nonce, unsigned long *hashes_done){
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
uint32_t hash64[8] __attribute__((aligned(32)));
uint32_t endiandata[32];
int kk=0;
#pragma unroll
for (; kk < 32; kk++){
be32enc(&endiandata[kk], ((uint32_t*)pdata)[kk]);
};
if (ptarget[7]==0) {
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFFFF)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFFF0)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFF00)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (((hash64[7]&0xFFFFF000)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFFFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (((hash64[7]&0xFFFF0000)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
Xhash(hash64, &endiandata);
if (fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;
}
bool scanhash_sse2_32(struct thr_info*thr, const unsigned char *pmidstate,
unsigned char *pdata,
unsigned char *phash1, unsigned char *phash,
const unsigned char *ptarget,
uint32_t max_nonce, uint32_t *last_nonce,
uint32_t nonce)
{
uint32_t *nNonce_p = (uint32_t *)(pdata + 76);
uint32_t m_midstate[8], m_w[16], m_w1[16];
__m128i m_4w[64] __attribute__ ((aligned (0x100)));
__m128i m_4hash[64] __attribute__ ((aligned (0x100)));
__m128i m_4hash1[64] __attribute__ ((aligned (0x100)));
__m128i offset;
int i;
pdata += 64;
/* Message expansion */
memcpy(m_midstate, pmidstate, sizeof(m_midstate));
memcpy(m_w, pdata, sizeof(m_w)); /* The 2nd half of the data */
memcpy(m_w1, phash1, sizeof(m_w1));
memset(m_4hash, 0, sizeof(m_4hash));
/* Transmongrify */
for (i = 0; i < 16; i++)
m_4w[i] = _mm_set1_epi32(m_w[i]);
for (i = 0; i < 16; i++)
m_4hash1[i] = _mm_set1_epi32(m_w1[i]);
for (i = 0; i < 64; i++)
sha256_consts_m128i[i] = _mm_set1_epi32(g_sha256_k[i]);
offset = _mm_set_epi32(0x3, 0x2, 0x1, 0x0);
for (;;)
{
int j;
m_4w[3] = _mm_add_epi32(offset, _mm_set1_epi32(nonce));
/* Some optimization can be done here W.R.T. precalculating some hash */
CalcSha256_x86 (m_4hash1, m_4w, m_midstate);
CalcSha256_x86 (m_4hash, m_4hash1, sha256_32init);
for (j = 0; j < 4; j++) {
if (unlikely(((uint32_t *)&(m_4hash[7]))[j] == 0)) {
/* We found a hit...so check it */
/* Use the C version for a check... */
for (i = 0; i < 8; i++) {
*(uint32_t *)&(phash)[i<<2] = ((uint32_t *)&(m_4hash[i]))[j];
}
if (fulltest(phash, ptarget)) {
nonce += j;
*last_nonce = nonce;
*nNonce_p = nonce;
return true;
}
}
}
if (unlikely((nonce >= max_nonce) || thr->work_restart)) {
*last_nonce = nonce;
return false;
}
nonce += 4;
}
}
int scanhash_scryptjane( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done)
{
scrypt_aligned_alloc YX, V;
uint8_t *X, *Y;
uint32_t N, chunk_bytes;
const uint32_t r = SCRYPT_R;
const uint32_t p = SCRYPT_P;
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t _ALIGN(64) endiandata[20];
const uint32_t first_nonce = pdata[19];
uint32_t nonce = first_nonce;
if (opt_benchmark)
ptarget[7] = 0x00ff;
for (int k = 0; k < 19; k++)
be32enc(&endiandata[k], pdata[k]);
//Nfactor = GetNfactor(data[17], ntime);
//if (Nfactor > scrypt_maxN) {
// return 1;
// //scrypt_fatal_error("scrypt: N out of range");
//}
N = (1 << ( opt_scrypt_n + 1));
chunk_bytes = SCRYPT_BLOCK_BYTES * r * 2;
if (!scrypt_alloc((uint64_t)N * chunk_bytes, &V)) return 1;
if (!scrypt_alloc((p + 1) * chunk_bytes, &YX)) {
scrypt_free(&V);
return 1;
}
Y = YX.ptr;
X = Y + chunk_bytes;
do {
const uint32_t Htarg = ptarget[7];
uint32_t hash[8];
be32enc(&endiandata[19], nonce);
scrypt_N_1_1((unsigned char *)endiandata, 80,
(unsigned char *)endiandata, 80,
N, (unsigned char *)hash, 32, X, Y, V.ptr);
if (hash[7] <= Htarg && fulltest(hash, ptarget)) {
pdata[19] = nonce;
*hashes_done = pdata[19] - first_nonce;
scrypt_free(&V);
scrypt_free(&YX);
return 1;
}
nonce++;
} while (nonce < max_nonce && !work_restart[thr_id].restart);
pdata[19] = nonce;
*hashes_done = pdata[19] - first_nonce + 1;
scrypt_free(&V);
scrypt_free(&YX);
return 0;
}
int scanhash_luffa(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
{
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
uint32_t hash64[8] __attribute__((aligned(32)));
uint32_t endiandata[32];
//char testdata[] = {"\x70\x00\x00\x00\x5d\x38\x5b\xa1\x14\xd0\x79\x97\x0b\x29\xa9\x41\x8f\xd0\x54\x9e\x7d\x68\xa9\x5c\x7f\x16\x86\x21\xa3\x14\x20\x10\x00\x00\x00\x00\x57\x85\x86\xd1\x49\xfd\x07\xb2\x2f\x3a\x8a\x34\x7c\x51\x6d\xe7\x05\x2f\x03\x4d\x2b\x76\xff\x68\xe0\xd6\xec\xff\x9b\x77\xa4\x54\x89\xe3\xfd\x51\x17\x32\x01\x1d\xf0\x73\x10\x00"};
//we need bigendian data...
//lessons learned: do NOT endianchange directly in pdata, this will all proof-of-works be considered as stale from minerd....
int kk=0;
for (; kk < 32; kk++)
{
be32enc(&endiandata[kk], ((uint32_t*)pdata)[kk]);
};
// if (opt_debug)
// {
// applog(LOG_DEBUG, "Thr: %02d, firstN: %08x, maxN: %08x, ToDo: %d", thr_id, first_nonce, max_nonce, max_nonce-first_nonce);
// }
/* I'm to lazy to put the loop in an inline function... so dirty copy'n'paste.... */
/* i know that i could set a variable, but i don't know how the compiler will optimize it, not that then the cpu needs to load the value *everytime* in a register */
if (ptarget[7]==0) {
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFFFF)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFFF0)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (((hash64[7]&0xFFFFFF00)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (((hash64[7]&0xFFFFF000)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else if (ptarget[7]<=0xFFFF)
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (((hash64[7]&0xFFFF0000)==0) &&
fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
else
{
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
luffahash(hash64, &endiandata);
if (fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
} while (n < max_nonce && !work_restart[thr_id].restart);
}
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;
}
