Skip to content

create seperate timers for cpu and gpu #1905

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

Open
wants to merge 6 commits into
base: develop
Choose a base branch
from
Open

create seperate timers for cpu and gpu #1905

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
896c0b1
create seperate timers for cpu and gpu
emre-safa Aug 1, 2025
d0ca8fe
modify the *solver*.stdout so that it contains both cpu and gpu for b…
emre-safa Aug 4, 2025
f908e6f
Apply suggestions from code review
emre-safa Aug 5, 2025
07ef983
change the order of cpu and gpu in components. Also add cpu and gpu u…
emre-safa Aug 5, 2025
e24aab3
add cpu/gpu timer for generation
emre-safa Aug 6, 2025
3dad637
update documentation
emre-safa Aug 7, 2025
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
43 changes: 34 additions & 9 deletions benchmark/solver/solver_common.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -476,6 +476,8 @@ struct SolverBenchmark : Benchmark<solver_benchmark_state<Generator>> {
for (auto stage : {"generate", "apply"}) {
solver_case[stage] = json::object();
solver_case[stage]["components"] = json::object();
solver_case[stage]["components"]["gpu"] = json::object();
solver_case[stage]["components"]["cpu"] = json::object();
}

IterationControl ic{timer};
Expand Down Expand Up @@ -529,19 +531,42 @@ struct SolverBenchmark : Benchmark<solver_benchmark_state<Generator>> {
}

{
auto apply_logger = create_operations_logger(
FLAGS_gpu_timer, FLAGS_nested_names, exec,
solver_case["apply"]["components"], 1);
exec->add_logger(apply_logger);
if (exec != exec->get_master()) {
exec->get_master()->add_logger(apply_logger);
std::shared_ptr<gko::log::ProfilerHook> apply_gpu_logger =
nullptr;
std::shared_ptr<gko::log::ProfilerHook> apply_cpu_logger =
nullptr;

if (FLAGS_gpu_timer) {
apply_gpu_logger = create_operations_logger(
true, FLAGS_nested_names, exec,
solver_case["apply"]["components"]["gpu"], 1);
}
if (FLAGS_cpu_timer) {
apply_cpu_logger = create_operations_logger(
false, FLAGS_nested_names, exec,
solver_case["apply"]["components"]["cpu"], 1);
}

if (FLAGS_cpu_timer) {
exec->add_logger(apply_cpu_logger);
if (exec != exec->get_master()) {
exec->get_master()->add_logger(apply_cpu_logger);
}
}
if (FLAGS_gpu_timer && exec != exec->get_master()) {
exec->add_logger(apply_gpu_logger);
}

solver->apply(state.b, x_clone);

exec->remove_logger(apply_logger);
if (exec != exec->get_master()) {
exec->get_master()->remove_logger(apply_logger);
if (FLAGS_cpu_timer) {
exec->remove_logger(apply_cpu_logger);
if (exec != exec->get_master()) {
exec->get_master()->remove_logger(apply_cpu_logger);
}
}
if (exec != exec->get_master() && FLAGS_gpu_timer) {
exec->remove_logger(apply_gpu_logger);
}
}

Expand Down
4 changes: 3 additions & 1 deletion benchmark/utils/timer.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: 2017 - 2024 The Ginkgo authors
// SPDX-FileCopyrightText: 2017 - 2025 The Ginkgo authors
//
// SPDX-License-Identifier: BSD-3-Clause

Expand All @@ -20,6 +20,8 @@ DEFINE_bool(gpu_timer, false,
"use gpu timer based on event. It is valid only when "
"executor is cuda or hip");

DEFINE_bool(cpu_timer, true, "use cpu timer based on event...");

DEFINE_string(
timer_method, "average",
"The statistical method for output of timer. Available options: "
Expand Down
Loading