AMReX_GpuLaunch.H File Reference

#include <AMReX_Config.H>
#include <AMReX_GpuQualifiers.H>
#include <AMReX_GpuKernelInfo.H>
#include <AMReX_GpuControl.H>
#include <AMReX_GpuTypes.H>
#include <AMReX_GpuError.H>
#include <AMReX_GpuRange.H>
#include <AMReX_GpuDevice.H>
#include <AMReX_GpuMemory.H>
#include <AMReX_GpuReduce.H>
#include <AMReX_Tuple.H>
#include <AMReX_Box.H>
#include <AMReX_Loop.H>
#include <AMReX_Extension.H>
#include <AMReX_BLassert.H>
#include <AMReX_TypeTraits.H>
#include <AMReX_GpuLaunchGlobal.H>
#include <AMReX_RandomEngine.H>
#include <AMReX_Algorithm.H>
#include <AMReX_Math.H>
#include <AMReX_Vector.H>
#include <cstddef>
#include <limits>
#include <algorithm>
#include <utility>
#include <AMReX_GpuLaunchMacrosG.H>
#include <AMReX_GpuLaunchFunctsG.H>
#include <AMReX_GpuLaunch.nolint.H>
#include <AMReX_CTOParallelForImpl.H>

Go to the source code of this file.

Classes
struct	amrex::Gpu::ComponentBox
struct	amrex::Gpu::GridSize
struct	amrex::Gpu::ExecutionConfig
struct	amrex::Gpu::ExecConfig

Namespaces
namespace	amrex
namespace	amrex::Gpu

Macros
#define	AMREX_GPU_NCELLS_PER_THREAD   3
#define	AMREX_GPU_Y_STRIDE   1
#define	AMREX_GPU_Z_STRIDE   1
#define	AMREX_LAUNCH_KERNEL(MT, blocks, threads, sharedMem, stream, ...)    amrex::launch_global<MT><<<blocks, threads, sharedMem, stream>>>(__VA_ARGS__)
#define	AMREX_LAUNCH_KERNEL_NOBOUND(blocks, threads, sharedMem, stream, ...)    amrex::launch_global <<<blocks, threads, sharedMem, stream>>>(__VA_ARGS__)

Functions
template<class L , class... Lambdas>
__global__ void	amrex::launch_global (L f0, Lambdas... fs)
template<class L >
__device__ void	amrex::call_device (L &&f0) noexcept
template<class L , class... Lambdas>
__device__ void	amrex::call_device (L &&f0, Lambdas &&... fs) noexcept
template<class L >
void	amrex::launch_host (L &&f0) noexcept
template<class L , class... Lambdas>
void	amrex::launch_host (L &&f0, Lambdas &&... fs) noexcept
constexpr std::size_t	amrex::Gpu::numThreadsPerBlockParallelFor ()
__host__ __device__ Box	amrex::Gpu::getThreadBox (const Box &bx, Long offset) noexcept
template<int MT>
ExecutionConfig	amrex::Gpu::makeExecutionConfig (Long N) noexcept
template<int MT>
ExecutionConfig	amrex::Gpu::makeExecutionConfig (const Box &box) noexcept
template<int MT>
Vector< ExecConfig >	amrex::Gpu::makeNExecutionConfigs (Long N) noexcept
template<int MT, int dim>
Vector< ExecConfig >	amrex::Gpu::makeNExecutionConfigs (BoxND< dim > const &box) noexcept
template<typename T , typename L , typename M = std::enable_if_t<std::is_integral_v<T>>>
void	amrex::ParallelForOMP (T n, L const &f) noexcept
template<typename L >
void	amrex::ParallelForOMP (Box const &box, L const &f) noexcept
template<typename T , typename L , typename M = std::enable_if_t<std::is_integral_v<T>>>
void	amrex::ParallelForOMP (Box const &box, T ncomp, L const &f) noexcept

Macro Definition Documentation

AMREX_GPU_NCELLS_PER_THREAD

#define AMREX_GPU_NCELLS_PER_THREAD   3

AMREX_GPU_Y_STRIDE

#define AMREX_GPU_Y_STRIDE   1

AMREX_GPU_Z_STRIDE

#define AMREX_GPU_Z_STRIDE   1

AMREX_LAUNCH_KERNEL

#define AMREX_LAUNCH_KERNEL	(		MT,
			blocks,
			threads,
			sharedMem,
			stream,
			...
	)		amrex::launch_global<MT><<<blocks, threads, sharedMem, stream>>>(__VA_ARGS__)

AMREX_LAUNCH_KERNEL_NOBOUND

#define AMREX_LAUNCH_KERNEL_NOBOUND	(		blocks,
			threads,
			sharedMem,
			stream,
			...
	)		amrex::launch_global <<<blocks, threads, sharedMem, stream>>>(__VA_ARGS__)