Tpetra parallel linear algebra
Version of the Day
|
Base class for the request (more or less a future) representing a pending nonblocking MPI operation. More...
#include <Tpetra_Details_iallreduce.hpp>
Public Member Functions | |
virtual | ~CommRequest () |
Destructor (virtual for memory safety of derived classes). More... | |
virtual void | wait ()=0 |
Wait on this communication request to complete. More... | |
virtual void | cancel ()=0 |
Cancel the pending communication request. More... | |
Base class for the request (more or less a future) representing a pending nonblocking MPI operation.
We recommend using auto
to interact with this class. For example, use auto
for the type of the return value of iallreduce() (see below).
Definition at line 89 of file Tpetra_Details_iallreduce.hpp.
|
inlinevirtual |
Destructor (virtual for memory safety of derived classes).
Definition at line 92 of file Tpetra_Details_iallreduce.hpp.
|
pure virtual |
Wait on this communication request to complete.
This is a blocking operation. The user is responsible for avoiding deadlock. This operation must also be idempotent.
Implemented in Tpetra::Details::Impl::IallreduceCommRequest< PacketType, SendLayoutType, SendDeviceType, RecvLayoutType, RecvDeviceType, 0 >, Tpetra::Details::Impl::IallreduceCommRequest< PacketType, SendLayoutType, SendDeviceType, RecvLayoutType, RecvDeviceType, 1 >, and Tpetra::Details::Impl::DeferredActionCommRequest.
|
pure virtual |
Cancel the pending communication request.
This operation must be idempotent.
Implemented in Tpetra::Details::Impl::IallreduceCommRequest< PacketType, SendLayoutType, SendDeviceType, RecvLayoutType, RecvDeviceType, 0 >, Tpetra::Details::Impl::IallreduceCommRequest< PacketType, SendLayoutType, SendDeviceType, RecvLayoutType, RecvDeviceType, 1 >, and Tpetra::Details::Impl::DeferredActionCommRequest.