Merge pull request #2570 from fireice-uk/xmr-stak-rx-dev

xmr-stak-rx 1.0.1 release

Author: psychocrypt

.appveyor.yml

@@ -1,5 +1,5 @@
 # image
-image: Visual Studio 2015
+image: Visual Studio 2017
 
 # build platform
 platform: x64
@@ -9,25 +9,26 @@ clone_folder: c:\xmr-stak
 
 install:
   - mkdir c:\xmr-stak-dep
-  - curl -sL https://github.com/fireice-uk/xmr-stak-dep/releases/download/v1/xmr-stak-dep.zip -o xmr-stak-dep.zip
-  - 7z x xmr-stak-dep.zip -o"c:\xmr-stak-dep" -y > nul
-  - appveyor DownloadFile  https://developer.nvidia.com/compute/cuda/8.0/prod/local_installers/cuda_8.0.44_windows-exe -FileName cuda_8.0.44_windows.exe
-  - cuda_8.0.44_windows.exe -s compiler_8.0 cudart_8.0 nvrtc_8.0 nvrtc_dev_8.0
-  - set PATH=%ProgramFiles%\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin;%ProgramFiles%\NVIDIA GPU Computing Toolkit\CUDA\v8.0\libnvvp;%PATH%
+  - curl -sL https://github.com/fireice-uk/xmr-stak-dep/releases/download/v2/xmr-stak-dep.zip -o xmr-stak-dep.zip
+  - 7z x xmr-stak-dep.zip -o"c:\" -y > nul
+  - dir c:\xmr-stak-dep
+  - appveyor DownloadFile "https://developer.nvidia.com/compute/cuda/10.1/Prod/network_installers/cuda_10.1.105_win10_network.exe" -FileName install_cuda.exe
+  - install_cuda.exe -s nvcc_10.1 cudart_10.1
+  - set PATH=%ProgramFiles%\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin;%ProgramFiles%\NVIDIA GPU Computing Toolkit\CUDA\v10.1\libnvvp;%PATH%
   - nvcc -V
 
 build_script:
-  - call "C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Tools\VsMSBuildCmd.bat"
+  - call "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build\vcvarsall.bat" x64 -vcvars_ver=14.11
   - cd c:\xmr-stak
   - mkdir build
   - cd build
   - set CMAKE_PREFIX_PATH=C:\xmr-stak-dep\hwloc;C:\xmr-stak-dep\libmicrohttpd;C:\xmr-stak-dep\openssl;
-  - cmake -G "Visual Studio 14 2015 Win64" -T v140,host=x64 .. -DCUDA_ARCH=30
+  - cmake -G "Visual Studio 15 2017 Win64" -T v141,host=x64 .. -DCUDA_ARCH=35
   - cmake --build . --config Release --target install
 
 test_script:
   - cd c:\xmr-stak\build\bin\Release
   - dir
   - copy C:\xmr-stak-dep\openssl\bin\* .
   - set XMRSTAK_NOWAIT=1
-  - xmr-stak.exe --help --noUAC
+  - xmr-stak-rx.exe --help --noUAC

.travis.yml

@@ -9,7 +9,7 @@ language: cpp
 cache:
   apt: true
   directories:
-    - $HOME/.cache/cuda
+    - $HOME/.cache/cuda9
 
 addons:
   apt:
@@ -21,25 +21,25 @@ addons:
 
 env:
   global:
-    - CUDA_ROOT: $HOME/.cache/cuda
+    - CUDA_ROOT: $HOME/.cache/cuda9
 
 matrix:
   include:
-    - os: linux
-      compiler: gcc
-      addons:
-        apt:
-          sources:
-            - ubuntu-toolchain-r-test
-          packages:
-            - *default_packages
-            - gcc-5
-            - g++-5
-      env:
-        - CMAKE_CXX_COMPILER=g++-5
-        - CMAKE_C_COMPILER=gcc-5
-        - XMRSTAK_CMAKE_FLAGS="-DCUDA_ARCH=30 -DOpenCL_ENABLE=OFF"
-
+#    - os: linux
+#      compiler: gcc
+#      addons:
+#        apt:
+#          sources:
+#            - ubuntu-toolchain-r-test
+#          packages:
+#            - *default_packages
+#            - gcc-5
+#            - g++-5
+#      env:
+#        - CMAKE_CXX_COMPILER=g++-5
+#        - CMAKE_C_COMPILER=gcc-5
+#        - XMRSTAK_CMAKE_FLAGS="-DCUDA_ARCH=30 -DOpenCL_ENABLE=OFF"
+#
     - os: linux
       compiler: gcc
       addons:
@@ -106,11 +106,11 @@ install:
       if [ $NVCC_FOUND -ne 0 ]; then
         mkdir -p $CUDA_ROOT &&
         cd $CUDA_ROOT &&
-        travis_retry wget https://developer.nvidia.com/compute/cuda/8.0/prod/local_installers/cuda_8.0.44_linux-run &&
+        travis_retry wget https://developer.nvidia.com/compute/cuda/9.0/Prod/local_installers/cuda_9.0.176_384.81_linux-run &&
         ls -la &&
         chmod u+x *-run &&
-        ./cuda_8.0.44_linux-run --silent --toolkit --toolkitpath=$CUDA_ROOT &&
-        rm -rf ./cuda_8.0.44_linux-run $CUDA_ROOT/{samples,jre,doc,share} &&
+        ./cuda_9.0.176_384.81_linux-run --silent --toolkit --toolkitpath=$CUDA_ROOT &&
+        rm -rf ./cuda_9.0.176_384.81_linux-run $CUDA_ROOT/{samples,jre,doc,share} &&
         cd -;
       fi
     fi;
@@ -123,4 +123,4 @@ script:
       cmake -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER} -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER} ${XMRSTAK_CMAKE_FLAGS} .;
     fi;
   - make VERBOSE=1 install
-  - ./bin/xmr-stak --help
+  - ./bin/xmr-stak-rx --help

xmrstak/backend/cpu/crypto/cryptonight_aesni.h

@@ -167,6 +167,7 @@ struct RandomX_generator
 		}
 		else if(algorithm_switched)
 		{
+			printer::inst()->print_msg(LDEBUG,"%s switched to %s",ctx[0]->last_algo.Name().c_str(), POW(ALGO).Name().c_str());
 			// remove old vm and re-initialize the full randomx context
 			for(size_t i = 0; i < N; i++)
 			{

xmrstak/backend/cpu/hwlocMemory.cpp

@@ -6,6 +6,24 @@
 
 #include <hwloc.h>
 
+static __hwloc_inline int
+xmrstak_set_membind_nodeset(hwloc_topology_t topology, hwloc_const_nodeset_t nodeset, hwloc_membind_policy_t policy, int flags)
+{
+#if HWLOC_API_VERSION >= 0x20000
+	return hwloc_set_membind(
+		topology,
+		nodeset,
+		policy,
+		flags| HWLOC_MEMBIND_BYNODESET);
+#else
+	return hwloc_set_membind_nodeset(
+		topology,
+		nodeset,
+		policy,
+		flags);
+#endif
+}
+
 /** pin memory to NUMA node
  *
  * Set the default memory policy for the current thread to bind memory to the
@@ -37,7 +55,7 @@ void bindMemoryToNUMANode(size_t puId)
 		hwloc_obj_t pu = hwloc_get_obj_by_depth(topology, depth, i);
 		if(pu->os_index == puId)
 		{
-			if(0 > hwloc_set_membind_nodeset(
+			if(0 > xmrstak_set_membind_nodeset(
 					   topology,
 					   pu->nodeset,
 					   HWLOC_MEMBIND_BIND,

xmrstak/backend/cpu/minethd.cpp

@@ -283,7 +283,7 @@ bool minethd::self_test()
 		xmrstak::globalStates::inst().iThreadCount = 1;
 		if(algo == POW(randomX))
 		{
-			printer::inst()->print_msg(L0, "start self test for 'randomx'");
+			printer::inst()->print_msg(L0, "start self test for 'randomx' (can be disabled with the command line option '--noTest')");
 			minethd::cn_on_new_job set_job;
 			func_multi_selector<1>(ctx, set_job, ::jconf::inst()->HaveHardwareAes(), algo);
 			miner_work work;
@@ -295,7 +295,7 @@ bool minethd::self_test()
 		}
 		else if(algo == POW(randomX_loki))
 		{
-			printer::inst()->print_msg(L0, "start self test for 'randomx_loki'");
+			printer::inst()->print_msg(L0, "start self test for 'randomx_loki' (can be disabled with the command line option '--noTest')");
 			minethd::cn_on_new_job set_job;
 			func_multi_selector<1>(ctx, set_job, ::jconf::inst()->HaveHardwareAes(), algo);
 			miner_work work;
@@ -307,7 +307,7 @@ bool minethd::self_test()
 		}
 		else if(algo == POW(randomX_wow))
 		{
-			printer::inst()->print_msg(L0, "start self test for 'randomx_wow'");
+			printer::inst()->print_msg(L0, "start self test for 'randomx_wow' (can be disabled with the command line option '--noTest')");
 			minethd::cn_on_new_job set_job;
 			func_multi_selector<1>(ctx, set_job, ::jconf::inst()->HaveHardwareAes(), algo);
 			miner_work work;
@@ -480,12 +480,6 @@ void minethd::func_multi_selector(cryptonight_ctx** ctx, minethd::cn_on_new_job&
 		on_new_job = nullptr;
 }
 
-void minethd::func_selector(cryptonight_ctx** ctx, bool bHaveAes, const xmrstak_algo& algo)
-{
-	minethd::cn_on_new_job dm;
-	func_multi_selector<1>(ctx, dm, bHaveAes, algo); // for testing us eauto, must be removed before the release
-}
-
 void minethd::work_main()
 {
 	multiway_work_main<1u>();

xmrstak/backend/cpu/minethd.hpp

@@ -24,7 +24,6 @@ class minethd : public iBackend
 
 	typedef void (*cn_on_new_job)(const miner_work&, cryptonight_ctx**);
 
-	static void func_selector(cryptonight_ctx**, bool bHaveAes, const xmrstak_algo& algo);
 	static bool thd_setaffinity(std::thread::native_handle_type h, uint64_t cpu_id);
 
 	static cryptonight_ctx* minethd_alloc_ctx();

xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu

@@ -322,6 +322,12 @@ extern "C" int cuda_get_deviceinfo(nvid_ctx* ctx)
 			hashMemSize = std::max(hashMemSize, algo.Mem());
 		}
 
+		const size_t dataset_size = getRandomXDatasetSize();
+		/* increase maxMemUsage by the dataset because the upper limits are
+		 * only for the scratchpad and does not take the randomX dataset into account.
+		 */
+		maxMemUsage += dataset_size;
+
 #ifdef WIN32
 		/* We use in windows bfactor (split slow kernel into smaller parts) to avoid
 		 * that windows is killing long running kernel.
@@ -346,7 +352,6 @@ extern "C" int cuda_get_deviceinfo(nvid_ctx* ctx)
 		size_t availableMem = freeMemory - (128u * byteToMiB) - 200u;
 		size_t limitedMemory = std::min(availableMem, maxMemUsage);
 
-		const size_t dataset_size = getRandomXDatasetSize();
 		if(limitedMemory <= dataset_size)
 			limitedMemory = 0;
 		else

xmrstak/cli/cli-miner.cpp

@@ -219,7 +219,7 @@ void do_guided_pool_config()
 	configTpl.set(std::string(tpl));
 	bool prompted = false;
 
-	auto& currency = params::inst().currency;
+	auto currency = params::inst().currency;
 	if(currency.empty() || !jconf::IsOnAlgoList(currency))
 	{
 		prompt_once(prompted);
@@ -236,7 +236,7 @@ void do_guided_pool_config()
 		currency = tmp;
 	}
 
-	auto& pool = params::inst().poolURL;
+	auto pool = params::inst().poolURL;
 	bool userSetPool = true;
 	if(pool.empty())
 	{
@@ -247,7 +247,7 @@ void do_guided_pool_config()
 		std::cin >> pool;
 	}
 
-	auto& userName = params::inst().poolUsername;
+	auto userName = params::inst().poolUsername;
 	if(userName.empty())
 	{
 		prompt_once(prompted);
@@ -257,7 +257,7 @@ void do_guided_pool_config()
 	}
 
 	bool stdin_flushed = false;
-	auto& passwd = params::inst().poolPasswd;
+	auto passwd = params::inst().poolPasswd;
 	if(passwd.empty() && !params::inst().userSetPwd)
 	{
 		prompt_once(prompted);
@@ -271,7 +271,7 @@ void do_guided_pool_config()
 		getline(std::cin, passwd);
 	}
 
-	auto& rigid = params::inst().poolRigid;
+	auto rigid = params::inst().poolRigid;
 	if(rigid.empty() && !params::inst().userSetRigid)
 	{
 		if(!use_simple_start())
@@ -366,7 +366,7 @@ void do_guided_config()
 	configTpl.set(std::string(tpl));
 	bool prompted = false;
 
-	auto& http_port = params::inst().httpd_port;
+	auto http_port = params::inst().httpd_port;
 	if(http_port == params::httpd_port_unset)
 	{
 		http_port = params::httpd_port_disabled;
@@ -943,6 +943,12 @@ int do_benchmark(int block_version, int wait_sec, int work_sec)
 
 	printer::inst()->print_msg(L0, "Prepare benchmark for block version %d", block_version);
 
+	if(block_version <= 0)
+	{
+		printer::inst()->print_msg(L0, "Block version must be >0, current value is %u.", block_version);
+		return 1;
+	}
+
 	uint8_t work[128];
 	memset(work, 0, 128);
 	work[0] = static_cast<uint8_t>(block_version);

xmrstak/jconf.cpp

@@ -100,10 +100,13 @@ configVal oConfigValues[] = {
 constexpr size_t iConfigCnt = (sizeof(oConfigValues) / sizeof(oConfigValues[0]));
 
 xmrstak::coin_selection coins[] = {
-	// name, userpool, devpool, default_pool_suggestion
+	// name, userpool, default_pool_suggestion
+	{"loki", {POW(randomX_loki)}, {POW(randomX_loki)}, nullptr},
+	{"monero", {POW(randomX)}, {POW(randomX)}, nullptr},
 	{"randomx", {POW(randomX)}, {POW(randomX)}, nullptr},
 	{"randomx_loki", {POW(randomX_loki)}, {POW(randomX_loki)}, nullptr},
-	{"randomx_wow", {POW(randomX_wow)}, {POW(randomX_wow)}, nullptr}
+	{"randomx_wow", {POW(randomX_wow)}, {POW(randomX_wow)}, nullptr},
+	{"wownero", {POW(randomX_wow)}, {POW(randomX_wow)}, nullptr}
 };
 
 constexpr size_t coin_algo_size = (sizeof(coins) / sizeof(coins[0]));

xmrstak/misc/executor.cpp

@@ -525,8 +525,9 @@ void executor::ex_main()
 			const char* rigid = params.userSetRigid ? params.poolRigid.c_str() : cfg.sRigId;
 			const char* pwd = params.userSetPwd ? params.poolPasswd.c_str() : cfg.sPasswd;
 			bool nicehash = cfg.nicehash || params.nicehashMode;
+			bool tls = params.poolUseTls;
 
-			pools.emplace_back(i, cfg.sPoolAddr, wallet, rigid, pwd, 9.9, params.poolUseTls, cfg.tls_fingerprint, nicehash);
+			pools.emplace_back(i, cfg.sPoolAddr, wallet, rigid, pwd, 9.9, tls, cfg.tls_fingerprint, nicehash);
 		}
 		else
 			pools.emplace_back(i, cfg.sPoolAddr, cfg.sWalletAddr, cfg.sRigId, cfg.sPasswd, cfg.weight, cfg.tls, cfg.tls_fingerprint, cfg.nicehash);