Represents how to schedule 1D, 2D, and 3D Cuda kernels. More...

#include <DeviceAdapterAlgorithmCuda.h>

Public Attributes
int	one_d_blocks

int	one_d_threads_per_block

int	two_d_blocks

dim3	two_d_threads_per_block

int	three_d_blocks

dim3	three_d_threads_per_block

Detailed Description

Represents how to schedule 1D, 2D, and 3D Cuda kernels.

ScheduleParameters represents how VTK-m should schedule different cuda kernel types. By default VTK-m uses a preset table based on the GPU's found at runtime.

When these defaults are insufficient for certain projects it is possible to override the defaults by using InitScheduleParameters.

Member Data Documentation

◆ one_d_blocks

int vtkm::cont::cuda::ScheduleParameters::one_d_blocks

◆ one_d_threads_per_block

int vtkm::cont::cuda::ScheduleParameters::one_d_threads_per_block

◆ three_d_blocks

int vtkm::cont::cuda::ScheduleParameters::three_d_blocks

◆ three_d_threads_per_block

dim3 vtkm::cont::cuda::ScheduleParameters::three_d_threads_per_block

◆ two_d_blocks

int vtkm::cont::cuda::ScheduleParameters::two_d_blocks

◆ two_d_threads_per_block

dim3 vtkm::cont::cuda::ScheduleParameters::two_d_threads_per_block

The documentation for this struct was generated from the following file:

DeviceAdapterAlgorithmCuda.h

Public Attributes