Browse Source

heavy: warnings and tabs (no changes)

master
Tanguy Pruvot 10 years ago
parent
commit
52dade56b5
  1. 179
      heavy/cuda_keccak512.cu
  2. 8
      heavy/heavy.cu

179
heavy/cuda_keccak512.cu

@ -23,120 +23,119 @@ __constant__ uint32_t c_PaddedMessage2[18]; // 44 bytes of remaining message (No @@ -23,120 +23,119 @@ __constant__ uint32_t c_PaddedMessage2[18]; // 44 bytes of remaining message (No
#include "cuda_helper.h"
#define U32TO64_LE(p) \
(((uint64_t)(*p)) | (((uint64_t)(*(p + 1))) << 32))
(((uint64_t)(*p)) | (((uint64_t)(*(p + 1))) << 32))
#define U64TO32_LE(p, v) \
*p = (uint32_t)((v)); *(p+1) = (uint32_t)((v) >> 32);
*p = (uint32_t)((v)); *(p+1) = (uint32_t)((v) >> 32);
static __device__ void mycpy72(uint32_t *d, const uint32_t *s) {
#pragma unroll 18
for (int k=0; k < 18; ++k) d[k] = s[k];
#pragma unroll 18
for (int k=0; k < 18; ++k) d[k] = s[k];
}
static __device__ void mycpy32(uint32_t *d, const uint32_t *s) {
#pragma unroll 8
for (int k=0; k < 8; ++k) d[k] = s[k];
#pragma unroll 8
for (int k=0; k < 8; ++k) d[k] = s[k];
}
typedef struct keccak_hash_state_t {
uint64_t state[25]; // 25*2
uint32_t buffer[72/4]; // 72
uint64_t state[25]; // 25*2
uint32_t buffer[72/4]; // 72
} keccak_hash_state;
__device__ void statecopy(uint64_t *d, uint64_t *s)
{
#pragma unroll 25
for (int i=0; i < 25; ++i)
d[i] = s[i];
#pragma unroll 25
for (int i=0; i < 25; ++i)
d[i] = s[i];
}
static const uint64_t host_keccak_round_constants[24] = {
0x0000000000000001ull, 0x0000000000008082ull,
0x800000000000808aull, 0x8000000080008000ull,
0x000000000000808bull, 0x0000000080000001ull,
0x8000000080008081ull, 0x8000000000008009ull,
0x000000000000008aull, 0x0000000000000088ull,
0x0000000080008009ull, 0x000000008000000aull,
0x000000008000808bull, 0x800000000000008bull,
0x8000000000008089ull, 0x8000000000008003ull,
0x8000000000008002ull, 0x8000000000000080ull,
0x000000000000800aull, 0x800000008000000aull,
0x8000000080008081ull, 0x8000000000008080ull,
0x0000000080000001ull, 0x8000000080008008ull
0x0000000000000001ull, 0x0000000000008082ull,
0x800000000000808aull, 0x8000000080008000ull,
0x000000000000808bull, 0x0000000080000001ull,
0x8000000080008081ull, 0x8000000000008009ull,
0x000000000000008aull, 0x0000000000000088ull,
0x0000000080008009ull, 0x000000008000000aull,
0x000000008000808bull, 0x800000000000008bull,
0x8000000000008089ull, 0x8000000000008003ull,
0x8000000000008002ull, 0x8000000000000080ull,
0x000000000000800aull, 0x800000008000000aull,
0x8000000080008081ull, 0x8000000000008080ull,
0x0000000080000001ull, 0x8000000080008008ull
};
__constant__ uint64_t c_keccak_round_constants[24];
__host__ __device__ void
keccak_block(uint64_t *s, const uint32_t *in, const uint64_t *keccak_round_constants) {
size_t i;
uint64_t t[5], u[5], v, w;
/* absorb input */
#pragma unroll 9
for (i = 0; i < 9 /* 72/8 */; i++, in += 2)
s[i] ^= U32TO64_LE(in);
for (i = 0; i < 24; i++) {
/* theta: c = a[0,i] ^ a[1,i] ^ .. a[4,i] */
t[0] = s[0] ^ s[5] ^ s[10] ^ s[15] ^ s[20];
t[1] = s[1] ^ s[6] ^ s[11] ^ s[16] ^ s[21];
t[2] = s[2] ^ s[7] ^ s[12] ^ s[17] ^ s[22];
t[3] = s[3] ^ s[8] ^ s[13] ^ s[18] ^ s[23];
t[4] = s[4] ^ s[9] ^ s[14] ^ s[19] ^ s[24];
/* theta: d[i] = c[i+4] ^ rotl(c[i+1],1) */
u[0] = t[4] ^ ROTL64(t[1], 1);
u[1] = t[0] ^ ROTL64(t[2], 1);
u[2] = t[1] ^ ROTL64(t[3], 1);
u[3] = t[2] ^ ROTL64(t[4], 1);
u[4] = t[3] ^ ROTL64(t[0], 1);
/* theta: a[0,i], a[1,i], .. a[4,i] ^= d[i] */
s[0] ^= u[0]; s[5] ^= u[0]; s[10] ^= u[0]; s[15] ^= u[0]; s[20] ^= u[0];
s[1] ^= u[1]; s[6] ^= u[1]; s[11] ^= u[1]; s[16] ^= u[1]; s[21] ^= u[1];
s[2] ^= u[2]; s[7] ^= u[2]; s[12] ^= u[2]; s[17] ^= u[2]; s[22] ^= u[2];
s[3] ^= u[3]; s[8] ^= u[3]; s[13] ^= u[3]; s[18] ^= u[3]; s[23] ^= u[3];
s[4] ^= u[4]; s[9] ^= u[4]; s[14] ^= u[4]; s[19] ^= u[4]; s[24] ^= u[4];
/* rho pi: b[..] = rotl(a[..], ..) */
v = s[ 1];
s[ 1] = ROTL64(s[ 6], 44);
s[ 6] = ROTL64(s[ 9], 20);
s[ 9] = ROTL64(s[22], 61);
s[22] = ROTL64(s[14], 39);
s[14] = ROTL64(s[20], 18);
s[20] = ROTL64(s[ 2], 62);
s[ 2] = ROTL64(s[12], 43);
s[12] = ROTL64(s[13], 25);
s[13] = ROTL64(s[19], 8);
s[19] = ROTL64(s[23], 56);
s[23] = ROTL64(s[15], 41);
s[15] = ROTL64(s[ 4], 27);
s[ 4] = ROTL64(s[24], 14);
s[24] = ROTL64(s[21], 2);
s[21] = ROTL64(s[ 8], 55);
s[ 8] = ROTL64(s[16], 45);
s[16] = ROTL64(s[ 5], 36);
s[ 5] = ROTL64(s[ 3], 28);
s[ 3] = ROTL64(s[18], 21);
s[18] = ROTL64(s[17], 15);
s[17] = ROTL64(s[11], 10);
s[11] = ROTL64(s[ 7], 6);
s[ 7] = ROTL64(s[10], 3);
s[10] = ROTL64( v, 1);
/* chi: a[i,j] ^= ~b[i,j+1] & b[i,j+2] */
v = s[ 0]; w = s[ 1]; s[ 0] ^= (~w) & s[ 2]; s[ 1] ^= (~s[ 2]) & s[ 3]; s[ 2] ^= (~s[ 3]) & s[ 4]; s[ 3] ^= (~s[ 4]) & v; s[ 4] ^= (~v) & w;
v = s[ 5]; w = s[ 6]; s[ 5] ^= (~w) & s[ 7]; s[ 6] ^= (~s[ 7]) & s[ 8]; s[ 7] ^= (~s[ 8]) & s[ 9]; s[ 8] ^= (~s[ 9]) & v; s[ 9] ^= (~v) & w;
v = s[10]; w = s[11]; s[10] ^= (~w) & s[12]; s[11] ^= (~s[12]) & s[13]; s[12] ^= (~s[13]) & s[14]; s[13] ^= (~s[14]) & v; s[14] ^= (~v) & w;
v = s[15]; w = s[16]; s[15] ^= (~w) & s[17]; s[16] ^= (~s[17]) & s[18]; s[17] ^= (~s[18]) & s[19]; s[18] ^= (~s[19]) & v; s[19] ^= (~v) & w;
v = s[20]; w = s[21]; s[20] ^= (~w) & s[22]; s[21] ^= (~s[22]) & s[23]; s[22] ^= (~s[23]) & s[24]; s[23] ^= (~s[24]) & v; s[24] ^= (~v) & w;
/* iota: a[0,0] ^= round constant */
s[0] ^= keccak_round_constants[i];
}
size_t i;
uint64_t t[5], u[5], v, w;
/* absorb input */
for (i = 0; i < 9 /* 72/8 */; i++, in += 2)
s[i] ^= U32TO64_LE(in);
for (i = 0; i < 24; i++) {
/* theta: c = a[0,i] ^ a[1,i] ^ .. a[4,i] */
t[0] = s[0] ^ s[5] ^ s[10] ^ s[15] ^ s[20];
t[1] = s[1] ^ s[6] ^ s[11] ^ s[16] ^ s[21];
t[2] = s[2] ^ s[7] ^ s[12] ^ s[17] ^ s[22];
t[3] = s[3] ^ s[8] ^ s[13] ^ s[18] ^ s[23];
t[4] = s[4] ^ s[9] ^ s[14] ^ s[19] ^ s[24];
/* theta: d[i] = c[i+4] ^ rotl(c[i+1],1) */
u[0] = t[4] ^ ROTL64(t[1], 1);
u[1] = t[0] ^ ROTL64(t[2], 1);
u[2] = t[1] ^ ROTL64(t[3], 1);
u[3] = t[2] ^ ROTL64(t[4], 1);
u[4] = t[3] ^ ROTL64(t[0], 1);
/* theta: a[0,i], a[1,i], .. a[4,i] ^= d[i] */
s[0] ^= u[0]; s[5] ^= u[0]; s[10] ^= u[0]; s[15] ^= u[0]; s[20] ^= u[0];
s[1] ^= u[1]; s[6] ^= u[1]; s[11] ^= u[1]; s[16] ^= u[1]; s[21] ^= u[1];
s[2] ^= u[2]; s[7] ^= u[2]; s[12] ^= u[2]; s[17] ^= u[2]; s[22] ^= u[2];
s[3] ^= u[3]; s[8] ^= u[3]; s[13] ^= u[3]; s[18] ^= u[3]; s[23] ^= u[3];
s[4] ^= u[4]; s[9] ^= u[4]; s[14] ^= u[4]; s[19] ^= u[4]; s[24] ^= u[4];
/* rho pi: b[..] = rotl(a[..], ..) */
v = s[ 1];
s[ 1] = ROTL64(s[ 6], 44);
s[ 6] = ROTL64(s[ 9], 20);
s[ 9] = ROTL64(s[22], 61);
s[22] = ROTL64(s[14], 39);
s[14] = ROTL64(s[20], 18);
s[20] = ROTL64(s[ 2], 62);
s[ 2] = ROTL64(s[12], 43);
s[12] = ROTL64(s[13], 25);
s[13] = ROTL64(s[19], 8);
s[19] = ROTL64(s[23], 56);
s[23] = ROTL64(s[15], 41);
s[15] = ROTL64(s[ 4], 27);
s[ 4] = ROTL64(s[24], 14);
s[24] = ROTL64(s[21], 2);
s[21] = ROTL64(s[ 8], 55);
s[ 8] = ROTL64(s[16], 45);
s[16] = ROTL64(s[ 5], 36);
s[ 5] = ROTL64(s[ 3], 28);
s[ 3] = ROTL64(s[18], 21);
s[18] = ROTL64(s[17], 15);
s[17] = ROTL64(s[11], 10);
s[11] = ROTL64(s[ 7], 6);
s[ 7] = ROTL64(s[10], 3);
s[10] = ROTL64( v, 1);
/* chi: a[i,j] ^= ~b[i,j+1] & b[i,j+2] */
v = s[ 0]; w = s[ 1]; s[ 0] ^= (~w) & s[ 2]; s[ 1] ^= (~s[ 2]) & s[ 3]; s[ 2] ^= (~s[ 3]) & s[ 4]; s[ 3] ^= (~s[ 4]) & v; s[ 4] ^= (~v) & w;
v = s[ 5]; w = s[ 6]; s[ 5] ^= (~w) & s[ 7]; s[ 6] ^= (~s[ 7]) & s[ 8]; s[ 7] ^= (~s[ 8]) & s[ 9]; s[ 8] ^= (~s[ 9]) & v; s[ 9] ^= (~v) & w;
v = s[10]; w = s[11]; s[10] ^= (~w) & s[12]; s[11] ^= (~s[12]) & s[13]; s[12] ^= (~s[13]) & s[14]; s[13] ^= (~s[14]) & v; s[14] ^= (~v) & w;
v = s[15]; w = s[16]; s[15] ^= (~w) & s[17]; s[16] ^= (~s[17]) & s[18]; s[17] ^= (~s[18]) & s[19]; s[18] ^= (~s[19]) & v; s[19] ^= (~v) & w;
v = s[20]; w = s[21]; s[20] ^= (~w) & s[22]; s[21] ^= (~s[22]) & s[23]; s[22] ^= (~s[23]) & s[24]; s[23] ^= (~s[24]) & v; s[24] ^= (~v) & w;
/* iota: a[0,0] ^= round constant */
s[0] ^= keccak_round_constants[i];
}
}
// Die Hash-Funktion

8
heavy/heavy.cu

@ -194,13 +194,13 @@ extern "C" void cuda_devicenames() @@ -194,13 +194,13 @@ extern "C" void cuda_devicenames()
// Can't be called directly in cpu-miner
extern "C" void cuda_devicereset()
{
cudaDeviceReset();
cudaDeviceReset();
}
static bool substringsearch(const char *haystack, const char *needle, int &match)
{
int hlen = strlen(haystack);
int nlen = strlen(needle);
int hlen = (int) strlen(haystack);
int nlen = (int) strlen(needle);
for (int i=0; i < hlen; ++i)
{
if (haystack[i] == ' ') continue;
@ -401,7 +401,7 @@ int scanhash_heavy_cpp(int thr_id, uint32_t *pdata, @@ -401,7 +401,7 @@ int scanhash_heavy_cpp(int thr_id, uint32_t *pdata,
if(actualNumberOfValuesInNonceVectorGPU > 0)
{
cudaMemcpy(cpu_nonceVector, d_nonceVector[thr_id], sizeof(uint32_t) * actualNumberOfValuesInNonceVectorGPU, cudaMemcpyDeviceToHost);
for (i=0; i<actualNumberOfValuesInNonceVectorGPU;++i)
{
uint32_t nonce = cpu_nonceVector[i];

Loading…
Cancel
Save