Commit d902c4a3 authored by gbalduzz's avatar gbalduzz
Browse files

Larger expansion order for test.

parent 09516106

include/dca/profiling/cuda_profiler.hpp

+1 −0
Original line number Diff line number Diff line
@@ -41,6 +41,7 @@ public: 
  static void stop_threading(int /*id*/){}

  template<class Concurrency>

  static void stop(Concurrency& /*conc*/, const std::string& /*name*/){}

  static void stop(const std::string& /*name*/){}



private:

  inline static bool active_ = false;

test/performance/phys/ctaux/bilayer_lattice_input.json

+1 −1
Original line number Diff line number Diff line 
{

  "physics": {

    "beta"                      :  4,

    "beta"                      :  10,

    "chemical-potential"        : 0

  },

test/performance/phys/ctaux/ctaux_walker_performance_test.cpp

+26 −13
Original line number Diff line number Diff line
@@ -14,9 +14,6 @@ 


#include <iostream>

#include <string>

#ifdef DCA_HAVE_CUDA

#include <cuda_profiler_api.h>

#endif



#include "dca/config/mc_options.hpp"

#include "dca/io/json/json_reader.hpp"
@@ -29,25 +26,37 @@ 
#include "dca/parallel/no_threading/no_threading.hpp"

#include "dca/phys/parameters/parameters.hpp"

#include "dca/profiling/events/time.hpp"

#include "dca/profiling/null_profiler.hpp"



#ifdef DCA_HAVE_CUDA

#include <cuda_profiler_api.h>

#include "dca/profiling/cuda_profiler.hpp"

#endif



const std::string input_dir = DCA_SOURCE_DIR "/test/performance/phys/ctaux/";



#ifdef DCA_HAVE_CUDA

constexpr dca::linalg::DeviceType device = dca::linalg::GPU;

const std::string device_name = "GPU";

using Profiler = dca::profiling::CudaProfiler;

#else

constexpr dca::linalg::DeviceType device = dca::linalg::CPU;

const std::string device_name = "CPU";

using Profiler = dca::profiling::NullProfiler;

#endif  // DCA_HAVE_CUDA



using RngType = dca::math::random::StdRandomWrapper<std::ranlux48_base>;

using Lattice = dca::phys::models::bilayer_lattice<dca::phys::domains::D4>;

using Model = dca::phys::models::TightBindingModel<Lattice>;

using Threading = dca::parallel::NoThreading;

using Concurrency = dca::parallel::NoConcurrency;

using Profiler = dca::profiling::CudaProfiler;

using Parameters = dca::phys::params::Parameters<Concurrency, Threading, Profiler, Model, RngType,

                                                 dca::phys::solver::CT_AUX>;

using Data = dca::phys::DcaData<Parameters>;

using Real = dca::config::McOptions::MCScalar;

template <dca::linalg::DeviceType device_t>

using Walker = dca::phys::solver::ctaux::CtauxWalker<device_t, Parameters, Data, Real>;

using Walker = dca::phys::solver::ctaux::CtauxWalker<device, Parameters, Data, Real>;



int main(int argc, char** argv) {

#ifdef DCA_HAVE_CUDA

  int submatrix_size = -1;

  int n_warmup = 30;

  int n_sweeps = 5;
@@ -92,15 +101,17 @@ int main(int argc, char** argv) { 
        walker.updateShell(i, n);

    }

  };

  std::cout << "\n\n  *********** GPU integration  ***************\n";

  std::cout << "\n\n  *********** " + device_name + " integration  ***************\n";

  std::cout << "Nr walkers: " << n_walkers << "\n\n";



#ifdef DCA_HAVE_CUDA

  dca::linalg::util::initializeMagma();

  dca::linalg::util::resizeHandleContainer(n_walkers);

#endif  // DCA_HAVE_CUDA



  RngType::resetCounter();

  std::vector<RngType> rngs;

  std::vector<Walker<dca::linalg::GPU>> walkers;

  std::vector<Walker> walkers;

  rngs.reserve(n_walkers);

  walkers.reserve(n_walkers);

  for (int i = 0; i < n_walkers; ++i) {
@@ -124,7 +135,9 @@ int main(int argc, char** argv) { 
  std::cout << "\n Warmed up.\n" << std::endl;



  // Timed section.

#ifdef DCA_HAVE_CUDA

  cudaProfilerStart();

#endif  // DCA_HAVE_CUDA

  Profiler::start();

  dca::profiling::WallTime start_t;
@@ -136,11 +149,11 @@ int main(int argc, char** argv) { 
    f.get();



  dca::profiling::WallTime integration_t;

  Profiler::stop();

  Profiler::stop("profile.txt");

#ifdef DCA_HAVE_CUDA

  cudaProfilerStop();

#endif  // DCA_HAVE_CUDA



  std::cout << std::endl;

  printTime("Integration GPU", start_t, integration_t);



#endif  // DCA_HAVE_CUDA

  printTime("Integration " + device_name, start_t, integration_t);

}