OpenCL GPU miner
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

716 lines
24 KiB

/*
* Copyright 2014 sgminer developers
*
* 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. See COPYING for more details.
*/
#include "algorithm.h"
#include "sph/sph_sha2.h"
#include "ocl.h"
#include "ocl/build_kernel.h"
#include "algorithm/scrypt.h"
#include "algorithm/animecoin.h"
#include "algorithm/inkcoin.h"
#include "algorithm/quarkcoin.h"
#include "algorithm/qubitcoin.h"
#include "algorithm/sifcoin.h"
#include "algorithm/darkcoin.h"
#include "algorithm/myriadcoin-groestl.h"
#include "algorithm/fuguecoin.h"
#include "algorithm/groestlcoin.h"
#include "algorithm/twecoin.h"
#include "algorithm/marucoin.h"
#include "algorithm/maxcoin.h"
11 years ago
#include "algorithm/talkcoin.h"
#include "algorithm/bitblock.h"
#include "algorithm/x14.h"
#include "compat.h"
#include <inttypes.h>
#include <string.h>
const char *algorithm_type_str[] = {
"Unknown",
"Scrypt",
"NScrypt",
"X11",
"X13",
"X14",
"X15",
"Keccak",
"Quarkcoin",
"Twecoin",
"Fugue256",
"NIST"
};
void sha256(const unsigned char *message, unsigned int len, unsigned char *digest)
{
sph_sha256_context ctx_sha2;
sph_sha256_init(&ctx_sha2);
sph_sha256(&ctx_sha2, message, len);
sph_sha256_close(&ctx_sha2, (void*)digest);
}
void gen_hash(const unsigned char *data, unsigned int len, unsigned char *hash)
{
unsigned char hash1[32];
sph_sha256_context ctx_sha2;
sph_sha256_init(&ctx_sha2);
sph_sha256(&ctx_sha2, data, len);
sph_sha256_close(&ctx_sha2, hash1);
sph_sha256(&ctx_sha2, hash1, 32);
sph_sha256_close(&ctx_sha2, hash);
}
#define CL_SET_BLKARG(blkvar) status |= clSetKernelArg(*kernel, num++, sizeof(uint), (void *)&blk->blkvar)
#define CL_SET_VARG(args, var) status |= clSetKernelArg(*kernel, num++, args * sizeof(uint), (void *)var)
#define CL_SET_ARG_N(n, var) do { status |= clSetKernelArg(*kernel, n, sizeof(var), (void *)&var); } while (0)
#define CL_SET_ARG_0(var) CL_SET_ARG_N(0, var)
#define CL_SET_ARG(var) CL_SET_ARG_N(num++, var)
#define CL_NEXTKERNEL_SET_ARG_N(n, var) do { kernel++; CL_SET_ARG_N(n, var); } while (0)
#define CL_NEXTKERNEL_SET_ARG_0(var) CL_NEXTKERNEL_SET_ARG_N(0, var)
#define CL_NEXTKERNEL_SET_ARG(var) CL_NEXTKERNEL_SET_ARG_N(num++, var)
static void append_scrypt_compiler_options(struct _build_kernel_data *data, struct cgpu_info *cgpu, struct _algorithm_t *algorithm)
{
char buf[255];
sprintf(buf, " -D LOOKUP_GAP=%d -D CONCURRENT_THREADS=%u -D NFACTOR=%d",
cgpu->lookup_gap, (unsigned int)cgpu->thread_concurrency, algorithm->nfactor);
strcat(data->compiler_options, buf);
sprintf(buf, "lg%utc%unf%u", cgpu->lookup_gap, (unsigned int)cgpu->thread_concurrency, algorithm->nfactor);
strcat(data->binary_filename, buf);
}
static void append_hamsi_compiler_options(struct _build_kernel_data *data, struct cgpu_info *cgpu, struct _algorithm_t *algorithm)
{
char buf[255];
sprintf(buf, " -D SPH_HAMSI_EXPAND_BIG=%d -D SPH_HAMSI_SHORT=%d ",
opt_hamsi_expand_big, ((opt_hamsi_short)?1:0));
strcat(data->compiler_options, buf);
sprintf(buf, "big%u%s", (unsigned int)opt_hamsi_expand_big, ((opt_hamsi_short)?"hs":""));
strcat(data->binary_filename, buf);
}
static cl_int queue_scrypt_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
unsigned char *midstate = blk->work->midstate;
cl_kernel *kernel = &clState->kernel;
unsigned int num = 0;
cl_uint le_target;
cl_int status = 0;
le_target = *(cl_uint *)(blk->work->device_target + 28);
memcpy(clState->cldata, blk->work->data, 80);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(clState->padbuffer8);
CL_SET_VARG(4, &midstate[0]);
CL_SET_VARG(4, &midstate[16]);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_maxcoin_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel = &clState->kernel;
unsigned int num = 0;
cl_int status = 0;
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->outputBuffer);
return status;
}
static cl_int queue_sph_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel = &clState->kernel;
unsigned int num = 0;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_darkcoin_mod_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// echo - search10
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_bitblock_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// echo - search10
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// hamsi - search11
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// fugue - search12
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// hamsi - search11
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// fugue - search12
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_bitblockold_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// combined echo, hamsi, fugue - shabal - whirlpool - search10
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_marucoin_mod_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// echo - search10
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// hamsi - search11
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// fugue - search12
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_marucoin_mod_old_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// combined echo, hamsi, fugue - search10
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
11 years ago
static cl_int queue_talkcoin_mod_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// groestl - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
11 years ago
// jh - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
11 years ago
// keccak - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
11 years ago
// skein - search4
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
11 years ago
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_x14_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// echo - search10
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// hamsi - search11
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// fugue - search12
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shabal - search13
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
static cl_int queue_x14_old_kernel(struct __clState *clState, struct _dev_blk_ctx *blk, __maybe_unused cl_uint threads)
{
cl_kernel *kernel;
unsigned int num;
cl_ulong le_target;
cl_int status = 0;
le_target = *(cl_ulong *)(blk->work->device_target + 24);
flip80(clState->cldata, blk->work->data);
status = clEnqueueWriteBuffer(clState->commandQueue, clState->CLbuffer0, true, 0, 80, clState->cldata, 0, NULL,NULL);
// blake - search
kernel = &clState->kernel;
num = 0;
CL_SET_ARG(clState->CLbuffer0);
CL_SET_ARG(clState->padbuffer8);
// bmw - search1
kernel = clState->extra_kernels;
CL_SET_ARG_0(clState->padbuffer8);
// groestl - search2
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// skein - search3
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// jh - search4
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// keccak - search5
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// luffa - search6
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// cubehash - search7
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// shavite - search8
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// simd - search9
CL_NEXTKERNEL_SET_ARG_0(clState->padbuffer8);
// combined echo, hamsi, fugue - shabal - search10
num = 0;
CL_NEXTKERNEL_SET_ARG(clState->padbuffer8);
CL_SET_ARG(clState->outputBuffer);
CL_SET_ARG(le_target);
return status;
}
typedef struct _algorithm_settings_t {
const char *name; /* Human-readable identifier */
algorithm_type_t type; //common algorithm type
double diff_multiplier1;
double diff_multiplier2;
double share_diff_multiplier;
uint32_t xintensity_shift;
uint32_t intensity_shift;
uint32_t found_idx;
unsigned long long diff_numerator;
uint32_t diff1targ;
size_t n_extra_kernels;
long rw_buffer_size;
cl_command_queue_properties cq_properties;
void (*regenhash)(struct work *);
cl_int (*queue_kernel)(struct __clState *, struct _dev_blk_ctx *, cl_uint);
void (*gen_hash)(const unsigned char *, unsigned int, unsigned char *);
void (*set_compile_options)(build_kernel_data *, struct cgpu_info *, algorithm_t *);
} algorithm_settings_t;
static algorithm_settings_t algos[] = {
// kernels starting from this will have difficulty calculated by using litecoin algorithm
#define A_SCRYPT(a) \
{ a, ALGO_SCRYPT, 1, 65536, 65536, 0, 0, 0xFF, 0xFFFFFFFFULL, 0x0000ffffUL, 0, -1, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, scrypt_regenhash, queue_scrypt_kernel, gen_hash, append_scrypt_compiler_options}
A_SCRYPT( "ckolivas" ),
A_SCRYPT( "alexkarnew" ),
A_SCRYPT( "alexkarnold" ),
A_SCRYPT( "bufius" ),
A_SCRYPT( "psw" ),
A_SCRYPT( "zuikkis" ),
#undef A_SCRYPT
// kernels starting from this will have difficulty calculated by using quarkcoin algorithm
#define A_QUARK(a, b) \
{ a, ALGO_QUARK, 256, 256, 256, 0, 0, 0xFF, 0xFFFFFFULL, 0x0000ffffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, b, queue_sph_kernel, gen_hash, NULL}
A_QUARK( "quarkcoin", quarkcoin_regenhash),
A_QUARK( "qubitcoin", qubitcoin_regenhash),
A_QUARK( "animecoin", animecoin_regenhash),
A_QUARK( "sifcoin", sifcoin_regenhash),
#undef A_QUARK
// kernels starting from this will have difficulty calculated by using bitcoin algorithm
#define A_DARK(a, b) \
{ a, ALGO_X11, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, b, queue_sph_kernel, gen_hash, NULL}
A_DARK( "darkcoin", darkcoin_regenhash),
A_DARK( "inkcoin", inkcoin_regenhash),
A_DARK( "myriadcoin-groestl", myriadcoin_groestl_regenhash),
#undef A_DARK
{ "twecoin", ALGO_TWE, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, twecoin_regenhash, queue_sph_kernel, sha256, NULL},
{ "maxcoin", ALGO_KECCAK, 1, 256, 1, 4, 15, 0x0F, 0xFFFFULL, 0x000000ffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, maxcoin_regenhash, queue_maxcoin_kernel, sha256, NULL},
{ "darkcoin-mod", ALGO_X11, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 10, 8 * 16 * 4194304, 0, darkcoin_regenhash, queue_darkcoin_mod_kernel, gen_hash, NULL},
{ "marucoin", ALGO_X13, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, marucoin_regenhash, queue_sph_kernel, gen_hash, append_hamsi_compiler_options},
{ "marucoin-mod", ALGO_X13, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 12, 8 * 16 * 4194304, 0, marucoin_regenhash, queue_marucoin_mod_kernel, gen_hash, append_hamsi_compiler_options},
{ "marucoin-modold", ALGO_X13, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 10, 8 * 16 * 4194304, 0, marucoin_regenhash, queue_marucoin_mod_old_kernel, gen_hash, append_hamsi_compiler_options},
{ "x14", ALGO_X14, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 13, 8 * 16 * 4194304, 0, x14_regenhash, queue_x14_kernel, gen_hash, append_hamsi_compiler_options},
{ "x14old", ALGO_X14, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 10, 8 * 16 * 4194304, 0, x14_regenhash, queue_x14_old_kernel, gen_hash, append_hamsi_compiler_options},
11 years ago
{ "bitblock", ALGO_X15, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 14, 4 * 16 * 4194304, 0, bitblock_regenhash, queue_bitblock_kernel, gen_hash, append_hamsi_compiler_options},
{ "bitblockold", ALGO_X15, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 10, 4 * 16 * 4194304, 0, bitblock_regenhash, queue_bitblockold_kernel, gen_hash, append_hamsi_compiler_options},
{ "talkcoin-mod", ALGO_NIST, 1, 1, 1, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 4, 8 * 16 * 4194304, 0, talkcoin_regenhash, queue_talkcoin_mod_kernel, gen_hash, NULL},
// kernels starting from this will have difficulty calculated by using fuguecoin algorithm
#define A_FUGUE(a, b) \
{ a, ALGO_FUGUE, 1, 256, 256, 0, 0, 0xFF, 0xFFFFULL, 0x0000ffffUL, 0, 0, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, b, queue_sph_kernel, sha256, NULL}
A_FUGUE( "fuguecoin", fuguecoin_regenhash),
A_FUGUE( "groestlcoin", groestlcoin_regenhash),
#undef A_FUGUE
// Terminator (do not remove)
{ NULL, ALGO_UNK, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NULL, NULL, NULL, NULL}
};
void copy_algorithm_settings(algorithm_t* dest, const char* algo)
{
algorithm_settings_t* src;
// Find algorithm settings and copy
for (src = algos; src->name; src++)
{
if (strcmp(src->name, algo) == 0)
{
strcpy(dest->name, src->name);
dest->type = src->type;
dest->diff_multiplier1 = src->diff_multiplier1;
dest->diff_multiplier2 = src->diff_multiplier2;
dest->share_diff_multiplier = src->share_diff_multiplier;
dest->xintensity_shift = src->xintensity_shift;
dest->intensity_shift = src->intensity_shift;
dest->found_idx = src->found_idx;
dest->diff_numerator = src->diff_numerator;
dest->diff1targ = src->diff1targ;
dest->n_extra_kernels = src->n_extra_kernels;
dest->rw_buffer_size = src->rw_buffer_size;
dest->cq_properties = src->cq_properties;
dest->regenhash = src->regenhash;
dest->queue_kernel = src->queue_kernel;
dest->gen_hash = src->gen_hash;
dest->set_compile_options = src->set_compile_options;
break;
}
}
// if not found
if (src->name == NULL)
{
applog(LOG_WARNING, "Algorithm %s not found, using %s.", algo, algos->name);
copy_algorithm_settings(dest, algos->name);
}
}
static const char *lookup_algorithm_alias(const char *lookup_alias, uint8_t *nfactor)
{
#define ALGO_ALIAS_NF(alias, name, nf) \
if (strcasecmp(alias, lookup_alias) == 0) { *nfactor = nf; return name; }
#define ALGO_ALIAS(alias, name) \
if (strcasecmp(alias, lookup_alias) == 0) return name;
ALGO_ALIAS_NF("scrypt", "ckolivas", 10);
ALGO_ALIAS_NF("scrypt", "ckolivas", 10);
ALGO_ALIAS_NF("adaptive-n-factor", "ckolivas", 11);
ALGO_ALIAS_NF("adaptive-nfactor", "ckolivas", 11);
ALGO_ALIAS_NF("nscrypt", "ckolivas", 11);
ALGO_ALIAS_NF("adaptive-nscrypt", "ckolivas", 11);
ALGO_ALIAS_NF("adaptive-n-scrypt", "ckolivas", 11);
ALGO_ALIAS("x11mod", "darkcoin-mod");
ALGO_ALIAS("x11", "darkcoin-mod");
ALGO_ALIAS("x13mod", "marucoin-mod");
ALGO_ALIAS("x13", "marucoin-mod");
ALGO_ALIAS("x13old", "marucoin-modold");
ALGO_ALIAS("x13modold", "marucoin-modold");
ALGO_ALIAS("x15mod", "bitblock");
ALGO_ALIAS("x15", "bitblock");
ALGO_ALIAS("x15modold", "bitblockold");
ALGO_ALIAS("x15old", "bitblockold");
ALGO_ALIAS("nist5", "talkcoin-mod");
ALGO_ALIAS("keccak", "maxcoin");
#undef ALGO_ALIAS
#undef ALGO_ALIAS_NF
return NULL;
}
void set_algorithm(algorithm_t* algo, const char* newname_alias)
{
const char* newname;
//load previous algorithm nfactor in case nfactor was applied before algorithm... or default to 10
uint8_t old_nfactor = ((algo->nfactor)?algo->nfactor:0);
uint8_t nfactor = 0;
if (!(newname = lookup_algorithm_alias(newname_alias, &nfactor)))
newname = newname_alias;
copy_algorithm_settings(algo, newname);
// use old nfactor if it was previously set and is different than the one set by alias
if ((old_nfactor > 0) && (old_nfactor != nfactor))
nfactor = old_nfactor;
// Doesn't matter for non-scrypt algorithms
if (nfactor > 0)
set_algorithm_nfactor(algo, nfactor);
}
void set_algorithm_nfactor(algorithm_t* algo, const uint8_t nfactor)
{
algo->nfactor = nfactor;
algo->n = (1 << nfactor);
//adjust algo type accordingly
switch (algo->type)
{
case ALGO_SCRYPT:
//if nfactor isnt 10, switch to NSCRYPT
if(algo->nfactor != 10)
algo->type = ALGO_NSCRYPT;
break;
//nscrypt
case ALGO_NSCRYPT:
//if nfactor is 10, switch to SCRYPT
if(algo->nfactor == 10)
algo->type = ALGO_SCRYPT;
break;
//ignore rest
default:
break;
}
}
bool cmp_algorithm(algorithm_t* algo1, algorithm_t* algo2)
{
return (strcmp(algo1->name, algo2->name) == 0) &&
(algo1->nfactor == algo2->nfactor);
}