Skip to content

Commit

Permalink
adjust hide communication unit tests
Browse files Browse the repository at this point in the history
  • Loading branch information
@omlins
omlins committed Aug 12, 2024
1 parent 3ec9329 commit 1b22a7f
Show file tree
Hide file tree
Showing 2 changed files with 8 additions and 8 deletions.
12 changes: 6 additions & 6 deletions test/ParallelKernel/test_parallel.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@ import Enzyme
@static if $package == $PKG_CUDA
call = @prettystring(1, @parallel f(A))
@test occursin("CUDA.@cuda blocks = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 32)) threads = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 32) stream = CUDA.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[3])))", call)
@test occursin("CUDA.synchronize(CUDA.stream())", call)
@test occursin("CUDA.synchronize(CUDA.stream(); blocking = true)", call)
call = @prettystring(1, @parallel ranges f(A))
@test occursin("CUDA.@cuda blocks = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 32)) threads = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 32) stream = CUDA.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ranges), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[3])))", call)
call = @prettystring(1, @parallel nblocks nthreads f(A))
Expand All @@ -46,7 +46,7 @@ import Enzyme
elseif $package == $PKG_AMDGPU
call = @prettystring(1, @parallel f(A))
@test occursin("AMDGPU.@roc gridsize = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 64)) groupsize = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 64) stream = AMDGPU.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[3])))", call)
@test occursin("AMDGPU.synchronize(AMDGPU.stream())", call)
@test occursin("AMDGPU.synchronize(AMDGPU.stream(); blocking = true)", call)
call = @prettystring(1, @parallel ranges f(A))
@test occursin("AMDGPU.@roc gridsize = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 64)) groupsize = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 64) stream = AMDGPU.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ranges), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[3])))", call)
call = @prettystring(1, @parallel nblocks nthreads f(A))
Expand Down Expand Up @@ -401,11 +401,11 @@ import Enzyme
end;
@testset "@synchronize" begin
@static if $package == $PKG_CUDA
@test @prettystring(1, @synchronize()) == "CUDA.synchronize()"
@test @prettystring(1, @synchronize(mystream)) == "CUDA.synchronize(mystream)"
@test @prettystring(1, @synchronize()) == "CUDA.synchronize(; blocking = true)"
@test @prettystring(1, @synchronize(mystream)) == "CUDA.synchronize(mystream; blocking = true)"
elseif $package == $PKG_AMDGPU
@test @prettystring(1, @synchronize()) == "AMDGPU.synchronize()"
@test @prettystring(1, @synchronize(mystream)) == "AMDGPU.synchronize(mystream)"
@test @prettystring(1, @synchronize()) == "AMDGPU.synchronize(; blocking = true)"
@test @prettystring(1, @synchronize(mystream)) == "AMDGPU.synchronize(mystream; blocking = true)"
end;
end;
@reset_parallel_kernel()
Expand Down
4 changes: 2 additions & 2 deletions test/test_parallel.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@ import ParallelStencil.@gorgeousexpand
@static if $package == $PKG_CUDA
call = @prettystring(1, @parallel f(A))
@test occursin("CUDA.@cuda blocks = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 32)) threads = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 32) stream = CUDA.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[3])))", call)
@test occursin("CUDA.synchronize(CUDA.stream())", call)
@test occursin("CUDA.synchronize(CUDA.stream(); blocking = true)", call)
call = @prettystring(1, @parallel ranges f(A))
@test occursin("CUDA.@cuda blocks = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 32)) threads = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 32) stream = CUDA.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ranges), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[3])))", call)
call = @prettystring(1, @parallel nblocks nthreads f(A))
Expand All @@ -45,7 +45,7 @@ import ParallelStencil.@gorgeousexpand
elseif $package == $PKG_AMDGPU
call = @prettystring(1, @parallel f(A))
@test occursin("AMDGPU.@roc gridsize = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 64)) groupsize = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A))); nthreads_x_max = 64) stream = AMDGPU.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ParallelStencil.ParallelKernel.get_ranges(A)))[3])))", call)
@test occursin("AMDGPU.synchronize(AMDGPU.stream())", call)
@test occursin("AMDGPU.synchronize(AMDGPU.stream(); blocking = true)", call)
call = @prettystring(1, @parallel ranges f(A))
@test occursin("AMDGPU.@roc gridsize = ParallelStencil.ParallelKernel.compute_nblocks(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)), ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 64)) groupsize = ParallelStencil.ParallelKernel.compute_nthreads(length.(ParallelStencil.ParallelKernel.promote_ranges(ranges)); nthreads_x_max = 64) stream = AMDGPU.stream() f(A, ParallelStencil.ParallelKernel.promote_ranges(ranges), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[1])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[2])), (Int64)(length((ParallelStencil.ParallelKernel.promote_ranges(ranges))[3])))", call)
call = @prettystring(1, @parallel nblocks nthreads f(A))
Expand Down

0 comments on commit 1b22a7f

Please sign in to comment.