Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature: make subdomain_m module fully asynchronous #167

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Feature: make subdomain_m module fully asynchronous #167

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 12 additions & 6 deletions example/time-paradigm.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@ program time_paradigm_m
character(len=:), allocatable :: steps_string, resolution_string
type(command_line_t) command_line
integer(int64) counter_start, counter_end, clock_rate
integer :: steps=200, resolution=64
integer :: steps=300, resolution=128

associate(me => this_image())
if (command_line%argument_present(["--help"])) then
Expand Down Expand Up @@ -56,16 +56,18 @@ function functional_programming_time() result(system_time)

call T%define(side=1., boundary_val=T_boundary, internal_val=T_internal_initial, n=resolution)

call system_clock(t_start_functional)

associate(dt => T%dx()*T%dy()/(4*alpha))
call system_clock(t_start_functional)

functional_programming: &
do step = 1, steps
sync all
T = T + dt * alpha * .laplacian. T
end do functional_programming

call system_clock(t_end_functional, clock_rate)
end associate

call system_clock(t_end_functional, clock_rate)
system_time = real(t_end_functional - t_start_functional)/real(clock_rate)

associate(L_infinity_norm => maxval(abs(T%values() - T_steady)))
Expand All @@ -80,16 +82,20 @@ function procedural_programming_time() result(system_time)
real system_time
type(subdomain_t) T

call T%define(side=1., boundary_val=0., internal_val=1., n=resolution)

associate(dt => T%dx()*T%dy()/(4*alpha))
call T%define(side=1., boundary_val=0., internal_val=1., n=resolution)
call system_clock(t_start_procedural)

procedural_programming: &
do step = 1, steps
sync all
call T%step(alpha*dt)
end do procedural_programming

call system_clock(t_end_procedural, clock_rate)
end associate

call system_clock(t_end_procedural, clock_rate)
system_time = real(t_end_procedural - t_start_procedural)/real(clock_rate)

associate(L_infinity_norm => maxval(abs(T%values() - T_steady)))
Expand Down
6 changes: 3 additions & 3 deletions src/matcha/subdomain_m.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,14 +14,14 @@ module subdomain_m
generic :: operator(.laplacian.) => laplacian
generic :: operator(*) => multiply
generic :: operator(+) => add
generic :: assignment(=) => assign_and_sync
generic :: assignment(=) => assign_
procedure dx
procedure dy
procedure dz
procedure values
procedure, private :: laplacian
procedure, private :: add
procedure, private :: assign_and_sync
procedure, private :: assign_
end type

interface
Expand Down Expand Up @@ -83,7 +83,7 @@ pure module function add(lhs, rhs) result(total)
type(subdomain_t) total
end function

module subroutine assign_and_sync(lhs, rhs)
module subroutine assign_(lhs, rhs)
implicit none
class(subdomain_t), intent(out) :: lhs
type(subdomain_t), intent(in) :: rhs
Expand Down
24 changes: 9 additions & 15 deletions src/matcha/subdomain_s.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,6 @@
allocate(halo_x(west:east, ny, nz)[*])
if (me>1) halo_x(east,:,:)[me-1] = self%s_(1,:,:)
if (me<num_subdomains) halo_x(west,:,:)[me+1] = self%s_(my_nx,:,:)
sync all
end procedure

module procedure dx
Expand Down Expand Up @@ -123,20 +122,23 @@
total%s_ = lhs%s_ + rhs%s_
end procedure

module procedure assign_and_sync
call assert(allocated(rhs%s_), "subdomain_t%assign_and_sync: allocated(rhs%s_)")
sync all
module procedure assign_
call assert(allocated(rhs%s_), "subdomain_t%assign_: allocated(rhs%s_)")
lhs%s_ = rhs%s_
if (me>1) halo_x(east,:,:)[me-1] = rhs%s_(1,:,:)
if (me<num_subdomains) halo_x(west,:,:)[me+1] = rhs%s_(my_nx,:,:)
sync all
call exchange_halo(rhs%s_)
end procedure

module procedure values
call assert(allocated(self%s_), "subdomain_t%values: allocated(self%s_)")
my_values = self%s_
end procedure

subroutine exchange_halo(s)
real, intent(in) :: s(:,:,:)
if (me>1) halo_x(east,:,:)[me-1] = s(1,:,:)
if (me<num_subdomains) halo_x(west,:,:)[me+1] = s(my_nx,:,:)
end subroutine

module procedure step

call assert(allocated(self%s_), "subdomain_t%laplacian: allocated(rhs%s_)")
Expand All @@ -149,10 +151,7 @@
call internal_points(increment)
call edge_points(increment)
call apply_boundary_condition(increment)

sync all
self%s_ = self%s_ + increment
sync all
call exchange_halo(self%s_)

contains
Expand Down Expand Up @@ -207,11 +206,6 @@ subroutine apply_boundary_condition(ds)
if (me==num_subdomains) ds(my_nx,:,:) = 0.
end subroutine

subroutine exchange_halo(s)
real, intent(in) :: s(:,:,:)
if (me>1) halo_x(east,:,:)[me-1] = s(1,:,:)
if (me<num_subdomains) halo_x(west,:,:)[me+1] = s(my_nx,:,:)
end subroutine

end procedure

Expand Down
17 changes: 11 additions & 6 deletions test/subdomain_test_m.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,8 @@ subroutine output(v)
critical
do j = 1, size(v,2)
do k = 1, size(v,3)
print *,"image ",this_image(),": ",j,k,v(:,j,k)
!print *,"image ",this_image(),": ",j,k,v(:,j,k)
print *,j,k,v(:,j,k)
end do
end do
end critical
Expand All @@ -65,7 +66,8 @@ function concave_laplacian() result(test_passes)
type(subdomain_t) f, laplacian_f
real, allocatable :: lap_f_vals(:,:,:)

call f%define(side=1., boundary_val=1., internal_val=2., n=21) ! internally constant subdomain with a step down at all surfaces
call f%define(side=1., boundary_val=1., internal_val=2., n=32) ! internally constant subdomain with a step down at all surfaces
sync all
laplacian_f = .laplacian. f
lap_f_vals = laplacian_f%values()

Expand Down Expand Up @@ -154,14 +156,15 @@ function concave_laplacian() result(test_passes)
function correct_steady_state() result(test_passes)
logical test_passes
type(subdomain_t) T
real, parameter :: T_boundary = 1., T_initial = 2., tolerance = 0.01, T_steady = T_boundary, alpha = 1.
integer, parameter :: steps = 6000
real, parameter :: T_boundary = 1., T_initial = 2., tolerance = 5.E-03, T_steady = T_boundary, alpha = 1.
integer, parameter :: steps = 25000
integer step

call T%define(side=1., boundary_val=T_boundary, internal_val=T_initial, n=21) ! const. internally with a step down at boundaries
call T%define(side=1., boundary_val=T_boundary, internal_val=T_initial, n=32) ! const. internally with a step down at boundaries

associate(dt => T%dx()*T%dy()*T%dz()/(4*alpha))
do step = 1, steps
sync all
T = T + dt * alpha * .laplacian. T
end do
end associate
Expand All @@ -174,7 +177,7 @@ function correct_steady_state() result(test_passes)
function functional_matches_procedural() result(test_passes)
logical test_passes
real, parameter :: tolerance = 1.E-06
integer, parameter :: steps = 6000, n=21
integer, parameter :: steps = 6000, n=32
real, parameter :: alpha = 1.
real, parameter :: side=1., boundary_val=1., internal_val=2.

Expand All @@ -195,6 +198,7 @@ function T_functional()

associate(dt => T%dx()*T%dy()/(4*alpha))
do step = 1, steps
sync all
T = T + dt * alpha * .laplacian. T
end do
end associate
Expand All @@ -211,6 +215,7 @@ function T_procedural()

associate(dt => T%dx()*T%dy()/(4*alpha))
do step = 1, steps
sync all
call T%step(alpha*dt)
end do
end associate
Expand Down