hmq1725 algo

Signed-off-by: Tanguy Pruvot <[email protected]>
DumaxFr · Mar 7, 2017 · 3d70026 · 3d70026
1 parent 4d95ef4
commit 3d70026
Show file tree

Hide file tree

Showing 14 changed files with 589 additions and 40 deletions.
diff --git a/Makefile.am b/Makefile.am
@@ -65,7 +65,7 @@ ccminer_SOURCES	= elist.h miner.h compat.h \
 			  x13/x13.cu x13/cuda_x13_hamsi512.cu x13/cuda_x13_fugue512.cu \
 			  x15/x14.cu x15/x15.cu x15/cuda_x14_shabal512.cu x15/cuda_x15_whirlpool.cu \
 			  x15/whirlpool.cu \
-			  x17/x17.cu x17/cuda_x17_haval256.cu x17/cuda_x17_sha512.cu \
+			  x17/x17.cu x17/hmq17.cu x17/cuda_x17_haval256.cu x17/cuda_x17_sha512.cu \
 			  x11/c11.cu x11/s3.cu x11/sib.cu x11/veltor.cu x11/cuda_streebog.cu
 
 # scrypt

diff --git a/algos.h b/algos.h
@@ -19,6 +19,7 @@ enum sha_algos {
 	ALGO_FUGUE256,		/* Fugue256 */
 	ALGO_GROESTL,
 	ALGO_HEAVY,		/* Heavycoin hash */
+	ALGO_HMQ1725,
 	ALGO_KECCAK,
 	ALGO_JACKPOT,
 	ALGO_LBRY,
@@ -73,6 +74,7 @@ static const char *algo_names[] = {
 	"fugue256",
 	"groestl",
 	"heavy",
+	"hmq1725",
 	"keccak",
 	"jackpot",
 	"lbry",
@@ -135,6 +137,8 @@ static inline int algo_to_int(char* arg)
 			i = ALGO_DMD_GR;
 		else if (!strcasecmp("doom", arg))
 			i = ALGO_LUFFA;
+		else if (!strcasecmp("hmq17", arg))
+			i = ALGO_HMQ1725;
 		else if (!strcasecmp("lyra2re", arg))
 			i = ALGO_LYRA2;
 		else if (!strcasecmp("lyra2rev2", arg))

diff --git a/bench.cpp b/bench.cpp
@@ -57,6 +57,7 @@ void algo_free_all(int thr_id)
 	free_fugue256(thr_id);
 	free_groestlcoin(thr_id);
 	free_heavy(thr_id);
+	free_hmq17(thr_id);
 	free_jackpot(thr_id);
 	free_lbry(thr_id);
 	free_luffa(thr_id);

diff --git a/ccminer.cpp b/ccminer.cpp
@@ -237,6 +237,7 @@ Options:\n\
 			fugue256    Fuguecoin\n\
 			groestl     Groestlcoin\n\
 			heavy       Heavycoin\n\
+			hmq1725     Doubloons / Espers\n\
 			jackpot     Jackpot\n\
 			keccak      Keccak-256 (Maxcoin)\n\
 			lbry        LBRY Credits (Sha/Ripemd)\n\
@@ -1600,6 +1601,7 @@ static bool stratum_gen_work(struct stratum_ctx *sctx, struct work *work)
 		opt_difficulty = 1.;
 
 	switch (opt_algo) {
+		case ALGO_HMQ1725: // should be 256 but... suprnova...
 		case ALGO_JACKPOT:
 		case ALGO_NEOSCRYPT:
 		case ALGO_SCRYPT:
@@ -2232,12 +2234,17 @@ static void *miner_thread(void *userdata)
 			rc = scanhash_myriad(thr_id, &work, max_nonce, &hashes_done);
 			break;
 
+		case ALGO_HMQ1725:
+			rc = scanhash_hmq17(thr_id, &work, max_nonce, &hashes_done);
+			break;
+
 		case ALGO_HEAVY:
 			rc = scanhash_heavy(thr_id, &work, max_nonce, &hashes_done, work.maxvote, HEAVYCOIN_BLKHDR_SZ);
 			break;
 		case ALGO_MJOLLNIR:
 			rc = scanhash_heavy(thr_id, &work, max_nonce, &hashes_done, 0, MNR_BLKHDR_SZ);
 			break;
+
 		case ALGO_KECCAK:
 			rc = scanhash_keccak256(thr_id, &work, max_nonce, &hashes_done);
 			break;

diff --git a/ccminer.vcxproj b/ccminer.vcxproj
@@ -544,6 +544,7 @@
     </CudaCompile>
     <CudaCompile Include="x15\cuda_x14_shabal512.cu" />
     <CudaCompile Include="x15\cuda_x15_whirlpool.cu" />
+    <CudaCompile Include="x17\hmq17.cu"/>
     <CudaCompile Include="x15\x15.cu" />
     <CudaCompile Include="x15\whirlpool.cu" />
     <CudaCompile Include="x17\x17.cu">

diff --git a/ccminer.vcxproj.filters b/ccminer.vcxproj.filters
@@ -703,6 +703,9 @@
     <CudaCompile Include="x17\cuda_x17_sha512.cu">
       <Filter>Source Files\CUDA\x17</Filter>
     </CudaCompile>
+    <CudaCompile Include="x17\hmq17.cu">
+      <Filter>Source Files\CUDA\x17</Filter>
+    </CudaCompile>
     <CudaCompile Include="x17\x17.cu">
       <Filter>Source Files\CUDA\x17</Filter>
     </CudaCompile>

diff --git a/miner.h b/miner.h
@@ -285,6 +285,7 @@ extern int scanhash_keccak256(int thr_id, struct work* work, uint32_t max_nonce,
 extern int scanhash_fresh(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done);
 extern int scanhash_fugue256(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done);
 extern int scanhash_groestlcoin(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done);
+extern int scanhash_hmq17(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done);
 extern int scanhash_heavy(int thr_id,struct work *work, uint32_t max_nonce, unsigned long *hashes_done, uint32_t maxvote, int blocklen);
 extern int scanhash_jackpot(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done);
 extern int scanhash_lbry(int thr_id, struct work *work, uint32_t max_nonce, unsigned long *hashes_done);
@@ -336,6 +337,7 @@ extern void free_fresh(int thr_id);
 extern void free_fugue256(int thr_id);
 extern void free_groestlcoin(int thr_id);
 extern void free_heavy(int thr_id);
+extern void free_hmq17(int thr_id);
 extern void free_jackpot(int thr_id);
 extern void free_lbry(int thr_id);
 extern void free_luffa(int thr_id);
@@ -852,6 +854,7 @@ void luffa_hash(void *state, const void *input);
 void fresh_hash(void *state, const void *input);
 void fugue256_hash(unsigned char* output, const unsigned char* input, int len);
 void heavycoin_hash(unsigned char* output, const unsigned char* input, int len);
+void hmq17hash(void *output, const void *input);
 void keccak256_hash(void *state, const void *input);
 unsigned int jackpothash(void *state, const void *input);
 void groestlhash(void *state, const void *input);

diff --git a/quark/cuda_bmw512.cu b/quark/cuda_bmw512.cu
@@ -1,13 +1,11 @@
 #include <stdio.h>
 #include <memory.h>
 
-#undef WANT_BMW512_80
+#define WANT_BMW512_80
 
 #include "cuda_helper.h"
 
-#ifdef WANT_BMW512_80
 __constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
-#endif
 
 #include "cuda_bmw512_sm3.cuh"
 
@@ -390,8 +388,6 @@ void quark_bmw512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t *
 	}
 }
 
-#ifdef WANT_BMW512_80
-
 __global__ __launch_bounds__(256, 2)
 void quark_bmw512_gpu_hash_80(uint32_t threads, uint32_t startNounce, uint64_t *g_hash)
 {
@@ -474,8 +470,6 @@ void quark_bmw512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce
 		quark_bmw512_gpu_hash_80_30<<<grid, block>>>(threads, startNounce, (uint64_t*)d_hash);
 }
 
-#endif
-
 __host__
 void quark_bmw512_cpu_init(int thr_id, uint32_t threads)
 {

diff --git a/quark/cuda_bmw512_sm3.cuh b/quark/cuda_bmw512_sm3.cuh
@@ -226,8 +226,6 @@ void quark_bmw512_gpu_hash_64_30(uint32_t threads, uint32_t startNounce, uint64_
 	}
 }
 
-#ifdef WANT_BMW512_80
-
 __global__
 void quark_bmw512_gpu_hash_80_30(uint32_t threads, uint32_t startNounce, uint64_t *g_hash)
 {
@@ -266,8 +264,6 @@ void quark_bmw512_gpu_hash_80_30(uint32_t threads, uint32_t startNounce, uint64_
 	}
 }
 
-#endif
-
 #else /* stripped stubs for other archs */
 __global__ void quark_bmw512_gpu_hash_64_30(uint32_t threads, uint32_t startNounce, uint64_t *g_hash, uint32_t *g_nonceVector) {}
 __global__ void quark_bmw512_gpu_hash_80_30(uint32_t threads, uint32_t startNounce, uint64_t *g_hash) {}

diff --git a/util.cpp b/util.cpp
@@ -2116,7 +2116,7 @@ void do_gpu_tests(void)
 
 	memset(work.data, 0, sizeof(work.data));
 	work.data[0] = 0;
-	scanhash_lbry(0, &work, 1, &done);
+	scanhash_hmq17(0, &work, 1, &done);
 
 	free(work_restart);
 	work_restart = NULL;
@@ -2179,6 +2179,9 @@ void print_hash_tests(void)
 	heavycoin_hash(&hash[0], &buf[0], 32);
 	printpfx("heavy", hash);
 
+	hmq17hash(&hash[0], &buf[0]);
+	printpfx("hmq1725", hash);
+
 	jackpothash(&hash[0], &buf[0]);
 	printpfx("jackpot", hash);
 

diff --git a/x17/cuda_x17_haval256.cu b/x17/cuda_x17_haval256.cu
@@ -259,14 +259,13 @@
 }
 
 __global__ /* __launch_bounds__(256, 6) */
-void x17_haval256_gpu_hash_64(const uint32_t threads, uint32_t startNounce, uint64_t *g_hash, uint32_t *g_nonceVector)
+void x17_haval256_gpu_hash_64(const uint32_t threads, uint64_t *g_hash, const int outlen)
 {
-	uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
+	const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
 	if (thread < threads)
 	{
-		uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
-		uint64_t hashPosition = nounce - startNounce;
-		uint64_t *pHash = &g_hash[hashPosition*8U];
+		const uint64_t hashPosition = thread*8U;
+		uint64_t *pHash = &g_hash[hashPosition];
 
 		uint32_t s0, s1, s2, s3, s4, s5, s6, s7;
 		const uint32_t u0 = s0 = 0x243F6A88;
@@ -288,7 +287,7 @@ void x17_haval256_gpu_hash_64(const uint32_t threads, uint32_t startNounce, uint
 			hash.h8[i] = pHash[i];
 		}
 
-///////// input big /////////////////////
+		///////// input big /////////////////////
 
 		uint32_t buf[32];
 
@@ -325,12 +324,13 @@ void x17_haval256_gpu_hash_64(const uint32_t threads, uint32_t startNounce, uint
 		pHash[1] = hash.h8[1];
 		pHash[2] = hash.h8[2];
 		pHash[3] = hash.h8[3];
-#ifdef NEED_HASH_512
-		pHash[4] = hash.h8[4];
-		pHash[5] = hash.h8[5];
-		pHash[6] = hash.h8[6];
-		pHash[7] = hash.h8[7];
-#endif
+
+		if (outlen == 512) {
+			pHash[4] = 0; //hash.h8[4];
+			pHash[5] = 0; //hash.h8[5];
+			pHash[6] = 0; //hash.h8[6];
+			pHash[7] = 0; //hash.h8[7];
+		}
 	}
 }
 
@@ -340,14 +340,12 @@ void x17_haval256_cpu_init(int thr_id, uint32_t threads)
 }
 
 __host__
-void x17_haval256_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order)
+void x17_haval256_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, const int outlen)
 {
 	const uint32_t threadsperblock = 256;
 
 	dim3 grid((threads + threadsperblock-1)/threadsperblock);
 	dim3 block(threadsperblock);
 
-	x17_haval256_gpu_hash_64 <<<grid, block>>> (threads, startNounce, (uint64_t*)d_hash, d_nonceVector);
-
-	//MyStreamSynchronize(NULL, order, thr_id);
+	x17_haval256_gpu_hash_64 <<<grid, block>>> (threads, (uint64_t*)d_hash, outlen);
 }
diff --git a/x17/cuda_x17_sha512.cu b/x17/cuda_x17_sha512.cu
@@ -89,13 +89,12 @@ uint64_t Tone(uint64_t* K, uint64_t* r, uint64_t* W, const int a, const int i)
 
 __global__
 /*__launch_bounds__(256, 4)*/
-void x17_sha512_gpu_hash_64(const uint32_t threads, uint32_t startNounce, uint64_t *g_hash, uint32_t *g_nonceVector)
+void x17_sha512_gpu_hash_64(const uint32_t threads, uint64_t *g_hash)
 {
 	const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
 	if (thread < threads)
 	{
-		uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
-		uint64_t hashPosition = nounce - startNounce;
+		const uint64_t hashPosition = thread;
 		uint64_t *pHash = &g_hash[hashPosition*8U];
 
 		uint64_t W[80];
@@ -161,14 +160,12 @@ void x17_sha512_cpu_init(int thr_id, uint32_t threads)
 }
 
 __host__
-void x17_sha512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order)
+void x17_sha512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash)
 {
 	const uint32_t threadsperblock = 256;
 
 	dim3 grid((threads + threadsperblock-1)/threadsperblock);
 	dim3 block(threadsperblock);
 
-	x17_sha512_gpu_hash_64 <<<grid, block>>> (threads, startNounce, (uint64_t*)d_hash, d_nonceVector);
-
-	//MyStreamSynchronize(NULL, order, thr_id);
+	x17_sha512_gpu_hash_64 <<<grid, block>>> (threads, (uint64_t*)d_hash);
 }