diff options
Diffstat (limited to 'xmrstak')
-rw-r--r-- | xmrstak/backend/nvidia/autoAdjust.hpp | 2 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/config.tpl | 10 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/jconf.cpp | 11 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/jconf.hpp | 1 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/minethd.cpp | 1 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp | 3 | ||||
-rw-r--r-- | xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu | 17 | ||||
-rw-r--r-- | xmrstak/cli/cli-miner.cpp | 57 |
8 files changed, 96 insertions, 6 deletions
diff --git a/xmrstak/backend/nvidia/autoAdjust.hpp b/xmrstak/backend/nvidia/autoAdjust.hpp index c6a7dca..be7d1ce 100644 --- a/xmrstak/backend/nvidia/autoAdjust.hpp +++ b/xmrstak/backend/nvidia/autoAdjust.hpp @@ -95,7 +95,7 @@ private: conf += std::string(" { \"index\" : ") + std::to_string(ctx.device_id) + ",\n" + " \"threads\" : " + std::to_string(ctx.device_threads) + ", \"blocks\" : " + std::to_string(ctx.device_blocks) + ",\n" + " \"bfactor\" : " + std::to_string(ctx.device_bfactor) + ", \"bsleep\" : " + std::to_string(ctx.device_bsleep) + ",\n" + - " \"affine_to_cpu\" : false,\n" + + " \"affine_to_cpu\" : false, \"sync_mode\" : 3,\n" + " },\n"; } } diff --git a/xmrstak/backend/nvidia/config.tpl b/xmrstak/backend/nvidia/config.tpl index 99dc023..5479172 100644 --- a/xmrstak/backend/nvidia/config.tpl +++ b/xmrstak/backend/nvidia/config.tpl @@ -9,6 +9,12 @@ R"===( * bsleep - Insert a delay of X microseconds between kernel launches. * Increase if you want to reduce GPU lag. Recommended setting on GUI systems - 100 * affine_to_cpu - This will affine the thread to a CPU. This can make a GPU miner play along nicer with a CPU miner. + * sync_mode - method used to synchronize the device + * documentation: http://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__DEVICE.html#group__CUDART__DEVICE_1g69e73c7dda3fc05306ae7c811a690fac + * 0 = cudaDeviceScheduleAuto + * 1 = cudaDeviceScheduleSpin - create a high load on one cpu thread per gpu + * 2 = cudaDeviceScheduleYield + * 3 = cudaDeviceScheduleBlockingSync (default) * * On the first run the miner will look at your system and suggest a basic configuration that will work, * you can try to tweak it from there to get the best performance. @@ -16,7 +22,9 @@ R"===( * A filled out configuration should look like this: * "gpu_threads_conf" : * [ - * { "index" : 0, "threads" : 17, "blocks" : 60, "bfactor" : 0, "bsleep" : 0, "affine_to_cpu" : false}, + * { "index" : 0, "threads" : 17, "blocks" : 60, "bfactor" : 0, "bsleep" : 0, + * "affine_to_cpu" : false, "sync_mode" : 3, + * }, * ], */ diff --git a/xmrstak/backend/nvidia/jconf.cpp b/xmrstak/backend/nvidia/jconf.cpp index 4208145..46c5726 100644 --- a/xmrstak/backend/nvidia/jconf.cpp +++ b/xmrstak/backend/nvidia/jconf.cpp @@ -123,16 +123,17 @@ bool jconf::GetGPUThreadConfig(size_t id, thd_cfg &cfg) if(!oThdConf.IsObject()) return false; - const Value *gid, *blocks, *threads, *bfactor, *bsleep, *aff; + const Value *gid, *blocks, *threads, *bfactor, *bsleep, *aff, *syncMode; gid = GetObjectMember(oThdConf, "index"); blocks = GetObjectMember(oThdConf, "blocks"); threads = GetObjectMember(oThdConf, "threads"); bfactor = GetObjectMember(oThdConf, "bfactor"); bsleep = GetObjectMember(oThdConf, "bsleep"); aff = GetObjectMember(oThdConf, "affine_to_cpu"); + syncMode = GetObjectMember(oThdConf, "sync_mode"); if(gid == nullptr || blocks == nullptr || threads == nullptr || - bfactor == nullptr || bsleep == nullptr || aff == nullptr) + bfactor == nullptr || bsleep == nullptr || aff == nullptr || syncMode == nullptr) { return false; } @@ -155,11 +156,17 @@ bool jconf::GetGPUThreadConfig(size_t id, thd_cfg &cfg) if(!aff->IsUint64() && !aff->IsBool()) return false; + if(!syncMode->IsNumber() || syncMode->GetInt() < 0 || syncMode->GetInt() > 3) + { + printer::inst()->print_msg(L0, "Error NVIDIA: sync_mode out of range or no number. ( range: 0 <= sync_mode < 4.)"); + return false; + } cfg.id = gid->GetInt(); cfg.blocks = blocks->GetInt(); cfg.threads = threads->GetInt(); cfg.bfactor = bfactor->GetInt(); cfg.bsleep = bsleep->GetInt(); + cfg.syncMode = syncMode->GetInt(); if(aff->IsNumber()) cfg.cpu_aff = aff->GetInt(); diff --git a/xmrstak/backend/nvidia/jconf.hpp b/xmrstak/backend/nvidia/jconf.hpp index b09a162..7f60f1d 100644 --- a/xmrstak/backend/nvidia/jconf.hpp +++ b/xmrstak/backend/nvidia/jconf.hpp @@ -28,6 +28,7 @@ public: bool bDoubleMode; bool bNoPrefetch; int32_t cpu_aff; + int syncMode; long long iCpuAff; }; diff --git a/xmrstak/backend/nvidia/minethd.cpp b/xmrstak/backend/nvidia/minethd.cpp index 9eab1c0..6e628fd 100644 --- a/xmrstak/backend/nvidia/minethd.cpp +++ b/xmrstak/backend/nvidia/minethd.cpp @@ -77,6 +77,7 @@ minethd::minethd(miner_work& pWork, size_t iNo, const jconf::thd_cfg& cfg) ctx.device_threads = (int)cfg.threads; ctx.device_bfactor = (int)cfg.bfactor; ctx.device_bsleep = (int)cfg.bsleep; + ctx.syncMode = cfg.syncMode; this->affinity = cfg.cpu_aff; std::unique_lock<std::mutex> lck(thd_aff_set); diff --git a/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp b/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp index 1b63379..afbdbaf 100644 --- a/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp +++ b/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp @@ -11,7 +11,8 @@ typedef struct { int device_blocks; int device_threads; int device_bfactor; - int device_bsleep; + int device_bsleep; + int syncMode; uint32_t *d_input; uint32_t inputlen; diff --git a/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu b/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu index bea2634..492201d 100644 --- a/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu +++ b/xmrstak/backend/nvidia/nvcc_code/cuda_extra.cu @@ -189,7 +189,22 @@ extern "C" int cryptonight_extra_cpu_init(nvid_ctx* ctx) } cudaDeviceReset(); - cudaSetDeviceFlags(cudaDeviceScheduleBlockingSync); + switch(ctx->syncMode) + { + case 0: + cudaSetDeviceFlags(cudaDeviceScheduleAuto); + break; + case 1: + cudaSetDeviceFlags(cudaDeviceScheduleSpin); + break; + case 2: + cudaSetDeviceFlags(cudaDeviceScheduleYield); + break; + case 3: + cudaSetDeviceFlags(cudaDeviceScheduleBlockingSync); + break; + + }; cudaDeviceSetCacheConfig(cudaFuncCachePreferL1); size_t hashMemSize; diff --git a/xmrstak/cli/cli-miner.cpp b/xmrstak/cli/cli-miner.cpp index 06acaea..546d226 100644 --- a/xmrstak/cli/cli-miner.cpp +++ b/xmrstak/cli/cli-miner.cpp @@ -66,6 +66,9 @@ void help() cout<<" -v, --version show version number"<<endl; cout<<" -V, --version-long show long version number"<<endl; cout<<" -c, --config FILE common miner configuration file"<<endl; +#ifdef _WIN32 + cout<<" --noUAC disable the UAC dialog"<<endl; +#endif #if (!defined(CONF_NO_AEON)) && (!defined(CONF_NO_MONERO)) cout<<" --currency NAME currency to mine: monero or aeon"<<endl; #endif @@ -269,6 +272,41 @@ void do_guided_config(bool userSetPasswd) std::cout<<"Configuration stored in file '"<<params::inst().configFile<<"'"<<std::endl; } +#ifdef _WIN32 +/** start the miner as administrator + * + * This function based on the stackoverflow post + * - source: https://stackoverflow.com/a/4893508 + * - author: Cody Gray + * - date: Feb 4 '11 + */ +void UACDialog(const std::string& binaryName, std::string& args) +{ + args += " --noUAC"; + SHELLEXECUTEINFO shExInfo = {0}; + shExInfo.cbSize = sizeof(shExInfo); + shExInfo.fMask = SEE_MASK_NOCLOSEPROCESS; + shExInfo.hwnd = 0; + shExInfo.lpVerb = "runas"; + shExInfo.lpFile = binaryName.c_str(); + // disable UAC dialog (else the miner will go into a infinite loop) + shExInfo.lpParameters = args.c_str(); + shExInfo.lpDirectory = 0; + shExInfo.nShow = SW_SHOW; + shExInfo.hInstApp = 0; + + if(ShellExecuteEx(&shExInfo)) + { + printer::inst()->print_msg(L0, + "This window has been opened because xmr-stak needed to run as administrator. It can be safely closed now."); + WaitForSingleObject(shExInfo.hProcess, INFINITE); + CloseHandle(shExInfo.hProcess); + // do not start the miner twice + std::exit(0); + } +} +#endif + int main(int argc, char *argv[]) { #ifndef CONF_NO_TLS @@ -302,6 +340,7 @@ int main(int argc, char *argv[]) } bool userSetPasswd = false; + bool uacDialog = true; for(int i = 1; i < argc; ++i) { std::string opName(argv[i]); @@ -424,6 +463,10 @@ int main(int argc, char *argv[]) } params::inst().configFile = argv[i]; } + else if(opName.compare("--noUAC") == 0) + { + uacDialog = false; + } else { printer::inst()->print_msg(L0, "Parameter unknown '%s'",argv[i]); @@ -432,6 +475,20 @@ int main(int argc, char *argv[]) } } +#ifdef _WIN32 + if(uacDialog) + { + std::string minerArgs; + for(int i = 1; i < argc; i++) + { + minerArgs += " "; + minerArgs += argv[i]; + } + + UACDialog(argv[0], minerArgs); + } +#endif + // check if we need a guided start if(!configEditor::file_exist(params::inst().configFile)) do_guided_config(userSetPasswd); |