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

Jump to bottom

Fix complex solver benchmark #1692

Open
wants to merge 5 commits into
base: develop
Choose a base branch
from
Open

Fix complex solver benchmark #1692

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
75ebae2
[ci] add benchmark tests with complex types
MarcelKoch Oct 11, 2024
4638128
[ci] add generated benchmark test reference
MarcelKoch Oct 11, 2024
4e6574d
[bench] fix residual norm logger for complex
MarcelKoch Oct 11, 2024
50a4605
fixup! [ci] add benchmark tests with complex types
MarcelKoch Oct 14, 2024
51e4fb7
review updates:
MarcelKoch Oct 23, 2024
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
2 changes: 1 addition & 1 deletion benchmark/test/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ function(add_benchmark_test test_name)
COMMAND ${Python3_EXECUTABLE} ${CMAKE_CURRENT_BINARY_DIR}/${test_name}.py $<TARGET_FILE:${test_name}> --generate
COMMENT "Regenerating reference output for ${test_name}"
WORKING_DIRECTORY "$<TARGET_FILE_DIR:ginkgo>")
add_dependencies(${regenerate_target} ${test_name})
add_dependencies(${regenerate_target} ${test_name} ${test_name}_dcomplex)
add_dependencies(benchmark_test_regenerate ${regenerate_target})
endfunction()
add_custom_target(benchmark_test_regenerate)
Expand Down
8 changes: 8 additions & 0 deletions benchmark/test/blas.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,3 +30,11 @@
expected_stdout="blas.profile.stdout",
expected_stderr="blas.profile.stderr",
)

# complex
test_framework.compare_output(
["-input", '[{"n": 100}]'],
expected_stdout="blas_dcomplex.simple.stdout",
expected_stderr="blas_dcomplex.simple.stderr",
use_complex=True
)
8 changes: 8 additions & 0 deletions benchmark/test/conversion.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,3 +67,11 @@
expected_stdout="conversion.profile.stdout",
expected_stderr="conversion.profile.stderr",
)

# complex
test_framework.compare_output(
["-input", '[{"size": 100, "stencil": "7pt"}]', "-formats", "coo,csr"],
expected_stdout="conversion_dcomplex.simple.stdout",
expected_stderr="conversion_dcomplex.simple.stderr",
use_complex=True
)
9 changes: 9 additions & 0 deletions benchmark/test/multi_vector_distributed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,3 +34,12 @@
expected_stderr="multi_vector_distributed.profile.stderr",
num_procs=3,
)

# complex
test_framework.compare_output_distributed(
["-input", '[{"n": 100}]'],
expected_stdout="multi_vector_distributed_dcomplex.simple.stdout",
expected_stderr="multi_vector_distributed_dcomplex.simple.stderr",
num_procs=3,
use_complex=True
)
8 changes: 8 additions & 0 deletions benchmark/test/preconditioner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,3 +66,11 @@
expected_stderr="preconditioner.reordered.stderr",
stdin='[{"size": 100, "stencil": "7pt"}]',
)

# complex
test_framework.compare_output(
["-input", '[{"size": 100, "stencil": "7pt"}]'],
expected_stdout="preconditioner_dcomplex.simple.stdout",
expected_stderr="preconditioner_dcomplex.simple.stderr",
use_complex=True
)
8 changes: 8 additions & 0 deletions benchmark/test/reference/blas_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
The operations are copy,axpy,scal
Running test case n = 100
Running blas: copy
Running blas: axpy
Running blas: scal
28 changes: 28 additions & 0 deletions benchmark/test/reference/blas_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
[
{
"n": 100,
"blas": {
"copy": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
},
"axpy": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
},
"scal": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
}
}
}
]
10 changes: 10 additions & 0 deletions benchmark/test/reference/conversion_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
The formats are coo,csr
Running test case stencil(100, 7pt)
Matrix is of size (125, 125), 725
Running conversion: coo-read
Running conversion: coo-csr
Running conversion: csr-read
Running conversion: csr-coo
31 changes: 31 additions & 0 deletions benchmark/test/reference/conversion_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
[
{
"size": 100,
"stencil": "7pt",
"conversion": {
"coo-read": {
"time": 1.0,
"repetitions": 10,
"completed": true
},
"coo-csr": {
"time": 1.0,
"repetitions": 10,
"completed": true
},
"csr-read": {
"time": 1.0,
"repetitions": 10,
"completed": true
},
"csr-coo": {
"time": 1.0,
"repetitions": 10,
"completed": true
}
},
"rows": 125,
"cols": 125,
"nonzeros": 725
}
]
8 changes: 8 additions & 0 deletions benchmark/test/reference/distributed_solver_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 1 running iterations
The random seed for right hand sides is 42
Running cg with 1000 iterations and residual goal of 1.000000e-06
The number of right hand sides is 1
Running test case stencil(100, 7pt, stencil)
Matrix is of size (125, 125)
Running solver: cg
59 changes: 59 additions & 0 deletions benchmark/test/reference/distributed_solver_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
[
{
"size": 100,
"stencil": "7pt",
"comm_pattern": "stencil",
"optimal": {
"spmv": "csr-csr"
},
"solver": {
"cg": {
"recurrent_residuals": [],
"true_residuals": [],
"implicit_residuals": [],
"iteration_timestamps": [],
"rhs_norm": 1.0,
"generate": {
"components": {
"generate(<typename>)": 1.0,
"free": 1.0,
"overhead": 1.0
},
"time": 1.0
},
"apply": {
"components": {
"apply(<typename>)": 1.0,
"iteration": 1.0,
"allocate": 1.0,
"dense::fill": 1.0,
"cg::initialize": 1.0,
"advanced_apply(<typename>)": 1.0,
"dense::row_gather": 1.0,
"csr::advanced_spmv": 1.0,
"dense::compute_squared_norm2": 1.0,
"dense::compute_sqrt": 1.0,
"copy(<typename>)": 1.0,
"dense::copy": 1.0,
"dense::compute_conj_dot_dispatch": 1.0,
"check(<typename>)": 1.0,
"residual_norm::residual_norm": 1.0,
"cg::step_1": 1.0,
"csr::spmv": 1.0,
"cg::step_2": 1.0,
"free": 1.0,
"overhead": 1.0
},
"iterations": 7,
"time": 1.0
},
"preconditioner": {},
"residual_norm": 1.0,
"repetitions": 1,
"completed": true
}
},
"rows": 125,
"cols": 125
}
]
8 changes: 8 additions & 0 deletions benchmark/test/reference/multi_vector_distributed_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
The operations are copy,axpy,scal
Running test case n = 100
Running blas: copy
Running blas: axpy
Running blas: scal
28 changes: 28 additions & 0 deletions benchmark/test/reference/multi_vector_distributed_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
[
{
"n": 100,
"blas": {
"copy": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
},
"axpy": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
},
"scal": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"completed": true
}
}
}
]
7 changes: 7 additions & 0 deletions benchmark/test/reference/preconditioner_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
Running with preconditioners: none
Running test case stencil(100, 7pt)
Matrix is of size (125, 125), 725
Running preconditioner: none
32 changes: 32 additions & 0 deletions benchmark/test/reference/preconditioner_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,32 @@
[
{
"size": 100,
"stencil": "7pt",
"preconditioner": {
"none": {
"generate": {
"components": {
"generate(<typename>)": 1.0,
"overhead": 1.0
},
"time": 1.0,
"repetitions": 10
},
"apply": {
"components": {
"apply(<typename>)": 1.0,
"copy(<typename>)": 1.0,
"dense::copy": 1.0,
"overhead": 1.0
},
"time": 1.0,
"repetitions": 10
},
"completed": true
}
},
"rows": 125,
"cols": 125,
"nonzeros": 725
}
]
8 changes: 8 additions & 0 deletions benchmark/test/reference/solver_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 1 running iterations
The random seed for right hand sides is 42
Running cg with 1000 iterations and residual goal of 1.000000e-06
The number of right hand sides is 1
Running test case stencil(100, 7pt)
Matrix is of size (125, 125)
Running solver: cg
56 changes: 56 additions & 0 deletions benchmark/test/reference/solver_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,56 @@
[
{
"size": 100,
"stencil": "7pt",
"optimal": {
"spmv": "csr"
},
"solver": {
"cg": {
"recurrent_residuals": [],
"true_residuals": [],
"implicit_residuals": [],
"iteration_timestamps": [],
"rhs_norm": 1.0,
"generate": {
"components": {
"generate(<typename>)": 1.0,
"free": 1.0,
"overhead": 1.0
},
"time": 1.0
},
"apply": {
"components": {
"apply(<typename>)": 1.0,
"iteration": 1.0,
"allocate": 1.0,
"dense::fill": 1.0,
"cg::initialize": 1.0,
"advanced_apply(<typename>)": 1.0,
"csr::advanced_spmv": 1.0,
"dense::compute_norm2_dispatch": 1.0,
"copy(<typename>)": 1.0,
"dense::copy": 1.0,
"dense::compute_conj_dot_dispatch": 1.0,
"check(<typename>)": 1.0,
"residual_norm::residual_norm": 1.0,
"cg::step_1": 1.0,
"csr::spmv": 1.0,
"cg::step_2": 1.0,
"free": 1.0,
"overhead": 1.0
},
"iterations": 7,
"time": 1.0
},
"preconditioner": {},
"residual_norm": 1.0,
"repetitions": 1,
"completed": true
}
},
"rows": 125,
"cols": 125
}
]
7 changes: 7 additions & 0 deletions benchmark/test/reference/sparse_blas_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
The operations are transpose
Running test case stencil(100, 7pt)
Matrix is of size (125, 125), 725
Running sparse_blas: transpose
25 changes: 25 additions & 0 deletions benchmark/test/reference/sparse_blas_dcomplex.simple.stdout
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,25 @@
[
{
"size": 100,
"stencil": "7pt",
"sparse_blas": {
"transpose": {
"time": 1.0,
"flops": 1.0,
"bandwidth": 1.0,
"repetitions": 10,
"components": {
"allocate": 1.0,
"components::fill_array": 1.0,
"csr::transpose": 1.0,
"free": 1.0,
"overhead": 1.0
},
"completed": true
}
},
"rows": 125,
"cols": 125,
"nonzeros": 725
}
]
8 changes: 8 additions & 0 deletions benchmark/test/reference/spmv_dcomplex.simple.stderr
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
Running on ReferenceExecutor
Running with 2 warm iterations and 10 running iterations
The random seed for right hand sides is 42
The formats are coo
The number of right hand sides is 1
Running test case stencil(100, 7pt)
Matrix is of size (125, 125), 725
Running spmv: coo
Loading
Loading