gs_device_mpi.F90 File Reference

Go to the source code of this file.

Data Types
type	gs_device_mpi::gs_device_mpi_buf_t
	Buffers for non-blocking communication and packing/unpacking. More...
type	gs_device_mpi::gs_device_mpi_t
	Gather-scatter communication using device MPI. The arrays are indexed per PE like send_pe and @ recv_pe. More...
interface	gs_device_mpi::hip_gs_pack
interface	gs_device_mpi::hip_gs_unpack
interface	gs_device_mpi::hip_gs_pack_vec
interface	gs_device_mpi::hip_gs_unpack_vec
interface	gs_device_mpi::device_mpi_init_reqs
interface	gs_device_mpi::device_mpi_free_reqs
interface	gs_device_mpi::device_mpi_isend
interface	gs_device_mpi::device_mpi_irecv
interface	gs_device_mpi::device_mpi_test
interface	gs_device_mpi::device_mpi_waitall
interface	gs_device_mpi::device_mpi_waitany

Modules
module	gs_device_mpi
	Defines GPU aware MPI gather-scatter communication.

Functions/Subroutines
subroutine	gs_device_mpi::gs_device_mpi_buf_init (this, pe_order, dof_stack, mark_dupes)
subroutine	gs_device_mpi::gs_device_mpi_buf_init_vec (this)
	Allocate this buffer's fused vector slab, sized for up to GS_VEC_NC components. Deferred to the first fused exchange, see gs_comm_t.
subroutine	gs_device_mpi::gs_device_mpi_buf_free (this)
subroutine	gs_device_mpi::gs_device_mpi_init (this, send_pe, recv_pe)
	Initialise MPI based communication method.
subroutine	gs_device_mpi::gs_device_mpi_init_vec (this)
	Allocate the fused vector send and receive slabs in device memory, sized for GS_VEC_NC components. Deferred to the first fused exchange, see gs_comm_t. The peer lists, dof maps and streams built in init are shared with the scalar exchange, so this stays rank local.
subroutine	gs_device_mpi::gs_device_mpi_free (this)
	Deallocate MPI based communication method.
subroutine	gs_device_mpi::gs_device_mpi_nbsend (this, u, n, tag, deps, strm)
	Post non-blocking send operations.
subroutine	gs_device_mpi::gs_device_mpi_nbrecv (this, tag)
	Post non-blocking receive operations.
subroutine	gs_device_mpi::gs_device_mpi_nbwait (this, u, n, op, strm)
	Wait for non-blocking operations.
subroutine	gs_device_mpi::gs_device_mpi_nbsend_vec (this, u, n, nc, tag, deps, strm)
	Fused nc-component send. u is the compact shared device buffer (component-outer, per-component stride n = nshared). Packs nc components into the interleaved send buffer and issues one Isend of nc*ndofs per peer.
subroutine	gs_device_mpi::gs_device_mpi_nbrecv_vec (this, tag, nc)
	Post non-blocking receives for a fused nc-component exchange.
subroutine	gs_device_mpi::gs_device_mpi_nbwait_vec (this, u, n, nc, op, strm)
	Wait for a fused nc-component exchange and unpack/reduce into u.