run_cuda Namespace Reference

Namespace with constants and functions to get CUDA device properties as well as default and test parallel parameters to use in CUDA implementation run optimization. More...

Functions
RunData	retrieveDeviceProperties (int num_device)

Variables
constexpr std::string_view	kOptimizeCUDADesc {"CUDA"}
constexpr std::string_view	kCUDAVersionHeader {"Cuda version"}
constexpr std::string_view	kCUDARuntimeHeader {"Cuda Runtime Version"}
constexpr std::array< unsigned int, 2 >	kParallelParamsDefault {{32, 4}}
	Default thread block dimensions (which is what parallel parameters corresponds to in CUDA implementation) More...
const std::set< std::array< unsigned int, 2 > >	kParallelParameterAltOptions
	Parallel parameter alternative options to run to retrieve optimized parallel parameters in CUDA implementation Parallel parameter corresponds to thread block dimensions in CUDA implementation OK to include default parallel parameters in alternative options but not required. More...

Detailed Description

Namespace with constants and functions to get CUDA device properties as well as default and test parallel parameters to use in CUDA implementation run optimization.

Function Documentation

retrieveDeviceProperties()

RunData run_cuda::retrieveDeviceProperties ( int  num_device )

inline

Definition at line 57 of file RunCUDASettings.h.

Here is the call graph for this function:

Here is the caller graph for this function:

Variable Documentation

kCUDARuntimeHeader

constexpr std::string_view run_cuda::kCUDARuntimeHeader {"Cuda Runtime Version"}

constexpr

Definition at line 55 of file RunCUDASettings.h.

kCUDAVersionHeader

constexpr std::string_view run_cuda::kCUDAVersionHeader {"Cuda version"}

constexpr

Definition at line 54 of file RunCUDASettings.h.

kOptimizeCUDADesc

constexpr std::string_view run_cuda::kOptimizeCUDADesc {"CUDA"}

constexpr

Definition at line 53 of file RunCUDASettings.h.

kParallelParameterAltOptions

const std::set<std::array<unsigned int, 2> > run_cuda::kParallelParameterAltOptions

Initial value:

{
  {16, 1}, {32, 1}, {32, 2}, {32, 3}, {32, 4}, {32, 5},
  {32, 6}, {32, 8}, {64, 1}, {64, 2}, {64, 3}, {64, 4},
  {128, 1}, {128, 2}, {256, 1}, {32, 10}, {32, 12}, {32, 14}, {32, 16},
  {64, 5}, {64, 6}, {64, 7}, {64, 8}, {128, 3}, {128, 4}, {256, 2}}

Parallel parameter alternative options to run to retrieve optimized parallel parameters in CUDA implementation Parallel parameter corresponds to thread block dimensions in CUDA implementation OK to include default parallel parameters in alternative options but not required.

Definition at line 105 of file RunCUDASettings.h.

kParallelParamsDefault

constexpr std::array<unsigned int, 2> run_cuda::kParallelParamsDefault {{32, 4}}

constexpr

Default thread block dimensions (which is what parallel parameters corresponds to in CUDA implementation)

Definition at line 82 of file RunCUDASettings.h.