hostallocator.h File Reference

#include <cstddef>
#include <memory>
#include "gromacs/math/paddedvector.h"
#include "gromacs/utility/alignedallocator.h"
#include "gromacs/utility/exceptions.h"

Include dependency graph for hostallocator.h:

This graph shows which files directly or indirectly include this file:

Description

Declares gmx::HostAllocationPolicy, gmx::HostAllocator, gmx::HostVector and gmx::PaddedHostVector, which are used to make/be standard library containers that can allocate memory suitable for transfers. Currently the only supported transfers using pinned memory are to CUDA GPUs, but other possibilities exist in future.

Todo:

This should not be in the public API, but it needs to be for the moment because state.h is in that API.

Author

Mark Abraham mark..nosp@m.j.ab.nosp@m.raham.nosp@m.@gma.nosp@m.il.co.nosp@m.m

Classes
class	gmx::HostAllocationPolicy
	Policy class for configuring gmx::Allocator, to manage allocations of memory that may be needed for e.g. GPU transfers. More...

Typedefs
template<class T >
using	gmx::HostAllocator = Allocator< T, HostAllocationPolicy >
	Memory allocator that uses HostAllocationPolicy. More...
template<class T >
using	gmx::HostVector = std::vector< T, HostAllocator< T >>
	Convenience alias for std::vector that uses HostAllocator.
template<class T >
using	gmx::PaddedHostVector = PaddedVector< T, HostAllocator< T >>
	Convenience alias for PaddedVector that uses HostAllocator.

Enumerations
enum	gmx::PinningPolicy : int { CannotBePinned, PinnedIfSupported }
	Helper enum for pinning policy of the allocation of HostAllocationPolicy. More...

Functions
template<class T1 , class T2 >
bool	gmx::operator== (const Allocator< T1, HostAllocationPolicy > &a, const Allocator< T2, HostAllocationPolicy > &b)
	Return true if two allocators are identical. More...
template<typename PinnableVector >
void	gmx::changePinningPolicy (PinnableVector *v, PinningPolicy pinningPolicy)
	Helper function for changing the pinning policy of a pinnable vector. More...