Gather-scatter communication using a crystal router on the device. More...

Inheritance diagram for gs_device_crystal::gs_device_crystal_t:

Collaboration diagram for gs_device_crystal::gs_device_crystal_t:

[legend]

Public Member Functions
procedure, pass(this)	init (this, send_pe, recv_pe)
	Initialise crystal router based device communication See gs_comm.f90 for details.

procedure, pass(this)	free (this)
	Deallocate crystal router based device communication.

procedure, pass(this)	nbsend (this, u, n, tag, deps, strm)
	Pack the shared vector and post the send of the first routing stage.

procedure, pass(this)	nbrecv (this, tag)
	Post the receives of the first routing stage.

procedure, pass(this)	nbwait (this, u, n, op, strm)
	Drive the remaining routing stages and reduce what is delivered into the shared vector.

procedure, pass(this)	nbsend_vec (this, u, n, nc, tag, deps, strm)
	Pack the shared vector and post the send of the first routing stage, fused nc-component.

procedure, pass(this)	nbrecv_vec (this, tag, nc)
	Post the receives of the first routing stage, fused nc-component.

procedure, pass(this)	nbwait_vec (this, u, n, nc, op, strm)
	Drive the remaining routing stages and reduce what is delivered into the shared vector, fused nc-component.

procedure(gs_comm_init), deferred, pass	init gs_comm_init

procedure(gs_comm_free), deferred, pass	free gs_comm_free

procedure(gs_nbsend), deferred, pass	nbsend gs_nbsend

procedure(gs_nbrecv), deferred, pass	nbrecv gs_nbrecv

procedure(gs_nbwait), deferred, pass	nbwait gs_nbwait

procedure, pass(this)	init_dofs (this)

procedure, pass(this)	free_dofs (this)

procedure, pass(this)	init_order (this, send_pe, recv_pe)
	Obtains which ranks to send and receive data from.

procedure, pass(this)	free_order (this)

procedure, pass(this)	take_schedule (this, src)
	Take over the gather-scatter schedule (dof lists and peer order) of src, avoiding a second (expensive) pass over the connectivity. The data is moved rather than copied, so src is left without a schedule and must not be used for communication afterwards (it can still be freed). No communication resources are set up here; complete the handover with init_schedule once src has been freed, so that the two backends never hold their resources at the same time.

procedure, pass(this)	init_schedule (this)
	Set up this communication method for the schedule taken over by take_schedule. Collective, as init is.

procedure, pass(this)	nbsend_vec (this, u, n, nc, tag, deps, strm)
	Fused vector halo exchange. Default implementations abort; backends that set vec_supported = .true. override them.

procedure, pass(this)	nbrecv_vec (this, tag, nc)
	Default fused vector receive. Abort unless a backend overrides it.

procedure, pass(this)	nbwait_vec (this, u, n, nc, op, strm)
	Default fused vector wait/reduce. Abort unless a backend overrides it.

Public Attributes
type(gs_crystal_plan_t)	plan
	The routing plan, worked out once from the schedule.

type(c_ptr), dimension(2)	buf_d = C_NULL_PTR
	Working buffer columns, plannwrk words each.

type(c_ptr)	sbuf_d = C_NULL_PTR
	Send buffer, plannsmax words.

type(c_ptr), dimension(2)	buf_v_d = C_NULL_PTR
	Fused vector working and send buffers, GS_VEC_NC times as long. Positions are interleaved, matching the gs vector pack kernels.

type(c_ptr)	sbuf_v_d = C_NULL_PTR

type(c_ptr), dimension(:), allocatable	keep_idx_d
	Per-stage index lists of the words that stay and the words that go, as positions in the stage's source column.

type(c_ptr), dimension(:), allocatable	send_idx_d

type(c_ptr), dimension(:), allocatable	keep_idx_v_d
	The same lists expanded over the components of the fused vector exchange, built on first use for the number of components seen.

type(c_ptr), dimension(:), allocatable	send_idx_v_d

type(c_ptr)	pack_keep_d = C_NULL_PTR
	Shared vector indices for the first stage and for the reduction.

type(c_ptr)	pack_send_d = C_NULL_PTR

type(c_ptr)	unpack_d = C_NULL_PTR

type(c_ptr)	sreqs = C_NULL_PTR
	Requests of the stage in flight, at most one send and two receives.

type(c_ptr)	rreqs = C_NULL_PTR

integer	nsreq = 0

integer	nrreq = 0

integer	tag = 0
	Tag of the operation in flight, taken from nbrecv since the later stages are driven from nbwait, which is not given one.

integer	vec_nc = 0
	Number of components the expanded index lists were built for, 0 until the first fused vector exchange.

type(stack_i4_t), dimension(:), allocatable	send_dof
	A list of stacks of dof indices local to this process to send to rank_i.

type(stack_i4_t), dimension(:), allocatable	recv_dof
	recv_dof(rank_i) is a stack of dof indices local to this process to receive from rank_i. size(recv_dof) == pe_size

integer, dimension(:), allocatable	send_pe
	Array of ranks that this process should send to.

integer, dimension(:), allocatable	recv_pe
	array of ranks that this process will receive messages from

logical	vec_supported = .false.
	Whether this backend implements the fused vector (multi-component) halo exchange (nbsend_vec/nbrecv_vec/nbwait_vec). When .false., the gs_op_r3 caller falls back to nc independent scalar exchanges.

Detailed Description

Definition at line 61 of file gs_device_crystal.F90.

Member Function/Subroutine Documentation

◆ free() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::free ( class(gs_device_crystal_t), intent(inout) this )

Definition at line 94 of file gs_device_crystal.F90.

◆ free() [2/2]

procedure(gs_comm_free), deferred, pass gs_comm::gs_comm_t::free

pure virtualinherited

Definition at line 71 of file gs_comm.f90.

◆ free_dofs()

procedure, pass(this) gs_comm::gs_comm_t::free_dofs ( class(gs_comm_t), intent(inout) this )

inherited

Definition at line 76 of file gs_comm.f90.

◆ free_order()

procedure, pass(this) gs_comm::gs_comm_t::free_order ( class(gs_comm_t), intent(inout) this )

inherited

Definition at line 78 of file gs_comm.f90.

◆ init() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::init	(	class(gs_device_crystal_t), intent(inout)	this,
		type(stack_i4_t), intent(inout)	send_pe,
		type(stack_i4_t), intent(inout)	recv_pe
	)

Definition at line 93 of file gs_device_crystal.F90.

◆ init() [2/2]

procedure(gs_comm_init), deferred, pass gs_comm::gs_comm_t::init

pure virtualinherited

Definition at line 70 of file gs_comm.f90.

◆ init_dofs()

procedure, pass(this) gs_comm::gs_comm_t::init_dofs ( class(gs_comm_t), intent(inout) this )

inherited

Definition at line 75 of file gs_comm.f90.

◆ init_order()

procedure, pass(this) gs_comm::gs_comm_t::init_order	(	class(gs_comm_t), intent(inout)	this,
		type(stack_i4_t), intent(inout)	send_pe,
		type(stack_i4_t), intent(inout)	recv_pe
	)

inherited

Parameters

send_pe,only	contains rank ids this porcesss should send to
recv_pe,only	the ranks this process should receive from

Definition at line 77 of file gs_comm.f90.

◆ init_schedule()

procedure, pass(this) gs_comm::gs_comm_t::init_schedule ( class(gs_comm_t), intent(inout) this )

inherited

Definition at line 80 of file gs_comm.f90.

◆ nbrecv() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbrecv	(	class(gs_device_crystal_t), intent(inout)	this,
		integer, intent(in)	tag
	)

Definition at line 96 of file gs_device_crystal.F90.

◆ nbrecv() [2/2]

procedure(gs_nbrecv), deferred, pass gs_comm::gs_comm_t::nbrecv

pure virtualinherited

Definition at line 73 of file gs_comm.f90.

◆ nbrecv_vec() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbrecv_vec	(	class(gs_device_crystal_t), intent(inout)	this,
		integer, intent(in)	tag,
		integer, intent(in)	nc
	)

Definition at line 99 of file gs_device_crystal.F90.

◆ nbrecv_vec() [2/2]

procedure, pass(this) gs_comm::gs_comm_t::nbrecv_vec	(	class(gs_comm_t), intent(inout)	this,
		integer, intent(in)	tag,
		integer, intent(in)	nc
	)

inherited

Definition at line 84 of file gs_comm.f90.

◆ nbsend() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbsend	(	class(gs_device_crystal_t), intent(inout)	this,
		real(kind=rp), dimension(n), intent(inout)	u,
		integer, intent(in)	n,
		integer, intent(in)	tag,
		type(c_ptr), intent(inout)	deps,
		type(c_ptr), intent(inout)	strm
	)

Definition at line 95 of file gs_device_crystal.F90.

◆ nbsend() [2/2]

procedure(gs_nbsend), deferred, pass gs_comm::gs_comm_t::nbsend

pure virtualinherited

Definition at line 72 of file gs_comm.f90.

◆ nbsend_vec() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbsend_vec	(	class(gs_device_crystal_t), intent(inout)	this,
		real(kind=rp), dimension(nc*n), intent(inout)	u,
		integer, intent(in)	n,
		integer, intent(in)	nc,
		integer, intent(in)	tag,
		type(c_ptr), intent(inout)	deps,
		type(c_ptr), intent(inout)	strm
	)

Parameters

u	compact shared device buffer, component-outer with stride n

Definition at line 98 of file gs_device_crystal.F90.

◆ nbsend_vec() [2/2]

procedure, pass(this) gs_comm::gs_comm_t::nbsend_vec	(	class(gs_comm_t), intent(inout)	this,
		real(kind=rp), dimension(nc*n), intent(inout)	u,
		integer, intent(in)	n,
		integer, intent(in)	nc,
		integer, intent(in)	tag,
		type(c_ptr), intent(inout)	deps,
		type(c_ptr), intent(inout)	strm
	)

inherited

Parameters

u	compact shared buffer, component-outer: u((c-1)*n + idx)
n	number of shared dofs (per component)
nc	number of components

Definition at line 83 of file gs_comm.f90.

◆ nbwait() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbwait	(	class(gs_device_crystal_t), intent(inout)	this,
		real(kind=rp), dimension(n), intent(inout)	u,
		integer, intent(in)	n,
		integer	op,
		type(c_ptr), intent(inout)	strm
	)

Definition at line 97 of file gs_device_crystal.F90.

◆ nbwait() [2/2]

procedure(gs_nbwait), deferred, pass gs_comm::gs_comm_t::nbwait

pure virtualinherited

Definition at line 74 of file gs_comm.f90.

◆ nbwait_vec() [1/2]

procedure, pass(this) gs_device_crystal::gs_device_crystal_t::nbwait_vec	(	class(gs_device_crystal_t), intent(inout)	this,
		real(kind=rp), dimension(nc*n), intent(inout)	u,
		integer, intent(in)	n,
		integer, intent(in)	nc,
		integer	op,
		type(c_ptr), intent(inout)	strm
	)

Definition at line 100 of file gs_device_crystal.F90.

◆ nbwait_vec() [2/2]

procedure, pass(this) gs_comm::gs_comm_t::nbwait_vec	(	class(gs_comm_t), intent(inout)	this,
		real(kind=rp), dimension(nc*n), intent(inout)	u,
		integer, intent(in)	n,
		integer, intent(in)	nc,
		integer	op,
		type(c_ptr), intent(inout)	strm
	)

inherited

Definition at line 85 of file gs_comm.f90.

◆ take_schedule()

procedure, pass(this) gs_comm::gs_comm_t::take_schedule	(	class(gs_comm_t), intent(inout)	this,
		class(gs_comm_t), intent(inout)	src
	)

inherited

Parameters

src	the backend to take the schedule from, left without one

Definition at line 79 of file gs_comm.f90.

Member Data Documentation

◆ buf_d

type(c_ptr), dimension(2) gs_device_crystal::gs_device_crystal_t::buf_d = C_NULL_PTR

Definition at line 65 of file gs_device_crystal.F90.

◆ buf_v_d

type(c_ptr), dimension(2) gs_device_crystal::gs_device_crystal_t::buf_v_d = C_NULL_PTR

Definition at line 70 of file gs_device_crystal.F90.

◆ keep_idx_d

type(c_ptr), dimension(:), allocatable gs_device_crystal::gs_device_crystal_t::keep_idx_d

Definition at line 74 of file gs_device_crystal.F90.

◆ keep_idx_v_d

type(c_ptr), dimension(:), allocatable gs_device_crystal::gs_device_crystal_t::keep_idx_v_d

Definition at line 77 of file gs_device_crystal.F90.

◆ nrreq

integer gs_device_crystal::gs_device_crystal_t::nrreq = 0

Definition at line 85 of file gs_device_crystal.F90.

◆ nsreq

integer gs_device_crystal::gs_device_crystal_t::nsreq = 0

Definition at line 85 of file gs_device_crystal.F90.

◆ pack_keep_d

type(c_ptr) gs_device_crystal::gs_device_crystal_t::pack_keep_d = C_NULL_PTR

Definition at line 79 of file gs_device_crystal.F90.

◆ pack_send_d

type(c_ptr) gs_device_crystal::gs_device_crystal_t::pack_send_d = C_NULL_PTR

Definition at line 80 of file gs_device_crystal.F90.

◆ plan

type(gs_crystal_plan_t) gs_device_crystal::gs_device_crystal_t::plan

Definition at line 63 of file gs_device_crystal.F90.

◆ recv_dof

type(stack_i4_t), dimension(:), allocatable gs_comm::gs_comm_t::recv_dof

inherited

Definition at line 58 of file gs_comm.f90.

◆ recv_pe

integer, dimension(:), allocatable gs_comm::gs_comm_t::recv_pe

inherited

Definition at line 64 of file gs_comm.f90.

◆ rreqs

type(c_ptr) gs_device_crystal::gs_device_crystal_t::rreqs = C_NULL_PTR

Definition at line 84 of file gs_device_crystal.F90.

◆ sbuf_d

type(c_ptr) gs_device_crystal::gs_device_crystal_t::sbuf_d = C_NULL_PTR

Definition at line 67 of file gs_device_crystal.F90.

◆ sbuf_v_d

type(c_ptr) gs_device_crystal::gs_device_crystal_t::sbuf_v_d = C_NULL_PTR

Definition at line 71 of file gs_device_crystal.F90.

◆ send_dof

type(stack_i4_t), dimension(:), allocatable gs_comm::gs_comm_t::send_dof

inherited

Definition at line 55 of file gs_comm.f90.

◆ send_idx_d

type(c_ptr), dimension(:), allocatable gs_device_crystal::gs_device_crystal_t::send_idx_d

Definition at line 74 of file gs_device_crystal.F90.

◆ send_idx_v_d

type(c_ptr), dimension(:), allocatable gs_device_crystal::gs_device_crystal_t::send_idx_v_d

Definition at line 77 of file gs_device_crystal.F90.

◆ send_pe

integer, dimension(:), allocatable gs_comm::gs_comm_t::send_pe

inherited

Note: : this will usually be fewer than the total number of ranks size(send_pe) <= pe_size

Definition at line 62 of file gs_comm.f90.

◆ sreqs

type(c_ptr) gs_device_crystal::gs_device_crystal_t::sreqs = C_NULL_PTR

Definition at line 83 of file gs_device_crystal.F90.

◆ tag

integer gs_device_crystal::gs_device_crystal_t::tag = 0

Definition at line 88 of file gs_device_crystal.F90.

◆ unpack_d

type(c_ptr) gs_device_crystal::gs_device_crystal_t::unpack_d = C_NULL_PTR

Definition at line 81 of file gs_device_crystal.F90.

◆ vec_nc

integer gs_device_crystal::gs_device_crystal_t::vec_nc = 0

Definition at line 91 of file gs_device_crystal.F90.

◆ vec_supported

logical gs_comm::gs_comm_t::vec_supported = .false.

inherited

Definition at line 68 of file gs_comm.f90.

The documentation for this type was generated from the following file:

/home/runner/work/neko/neko/src/gs/bcknd/device/gs_device_crystal.F90

Public Member Functions

Public Attributes

Detailed Description

Member Function/Subroutine Documentation

◆ free() [1/2]

◆ free() [2/2]

◆ free_dofs()

◆ free_order()

◆ init() [1/2]

◆ init() [2/2]

◆ init_dofs()

◆ init_order()

◆ init_schedule()

◆ nbrecv() [1/2]

◆ nbrecv() [2/2]

◆ nbrecv_vec() [1/2]

◆ nbrecv_vec() [2/2]

◆ nbsend() [1/2]

◆ nbsend() [2/2]

◆ nbsend_vec() [1/2]

◆ nbsend_vec() [2/2]

◆ nbwait() [1/2]

◆ nbwait() [2/2]

◆ nbwait_vec() [1/2]

◆ nbwait_vec() [2/2]

◆ take_schedule()

Member Data Documentation

◆ buf_d

◆ buf_v_d

◆ keep_idx_d

◆ keep_idx_v_d

◆ nrreq

◆ nsreq

◆ pack_keep_d

◆ pack_send_d

◆ plan

◆ recv_dof

◆ recv_pe

◆ rreqs

◆ sbuf_d

◆ sbuf_v_d

◆ send_dof

◆ send_idx_d

◆ send_idx_v_d

◆ send_pe

◆ sreqs

◆ tag

◆ unpack_d

◆ vec_nc

◆ vec_supported