Neko
0.8.1
A portable framework for high-order spectral element flow simulations
|
Defines GPU aware MPI gather-scatter communication. More...
Data Types | |
type | gs_device_mpi_buf_t |
Buffers for non-blocking communication and packing/unpacking. More... | |
type | gs_device_mpi_t |
Gather-scatter communication using device MPI. The arrays are indexed per PE like send_pe and @ recv_pe. More... | |
interface | hip_gs_pack |
interface | hip_gs_unpack |
interface | device_mpi_init_reqs |
interface | device_mpi_free_reqs |
interface | device_mpi_isend |
interface | device_mpi_irecv |
interface | device_mpi_test |
interface | device_mpi_waitall |
interface | device_mpi_waitany |
Functions/Subroutines | |
subroutine | gs_device_mpi_buf_init (this, pe_order, dof_stack, mark_dupes) |
subroutine | gs_device_mpi_buf_free (this) |
subroutine | gs_device_mpi_init (this, send_pe, recv_pe) |
Initialise MPI based communication method. More... | |
subroutine | gs_device_mpi_free (this) |
Deallocate MPI based communication method. More... | |
subroutine | gs_device_mpi_nbsend (this, u, n, deps, strm) |
Post non-blocking send operations. More... | |
subroutine | gs_device_mpi_nbrecv (this) |
Post non-blocking receive operations. More... | |
subroutine | gs_device_mpi_nbwait (this, u, n, op, strm) |
Wait for non-blocking operations. More... | |
Defines GPU aware MPI gather-scatter communication.
|
private |
|
private |
|
private |
Deallocate MPI based communication method.
Definition at line 304 of file gs_device_mpi.F90.
|
private |
Initialise MPI based communication method.
Definition at line 274 of file gs_device_mpi.F90.
|
private |
Post non-blocking receive operations.
Definition at line 400 of file gs_device_mpi.F90.
|
private |
Post non-blocking send operations.
Definition at line 326 of file gs_device_mpi.F90.
|
private |
Wait for non-blocking operations.
Definition at line 413 of file gs_device_mpi.F90.