IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Extend throughput benchmark with device CLI parameter (#20816)

Browse Source 
* Extend throughput benchmark CLI parameters

* Added device name as the second CLI parameter with default CPU value

* Update samples/cpp/benchmark/throughput_benchmark/main.cpp

Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>

* Fix comments to review

* Modified python version
* Modified documentation

* Fix comments to review

* Fixed the comment
* Modified python doc
* Fixed device name handling in python version

* Update main.cpp

* Update throughput_benchmark.py

---------

Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
This commit is contained in:
Anatoliy Talamanov 2023-11-03 08:57:04 +00:00 committed by Alexander Nesterov
parent cf3d8435bd
commit e6e7ee5e6f
4 changed files with 17 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/articles_en/learn_openvino/openvino_samples/cpp_sample_throughput_benchmark.md
View File

@ -75,7 +75,7 @@ Running
.. code-block:: sh .. code-block:: sh
throughput_benchmark <path_to_model> throughput_benchmark <path_to_model> <device_name>(default: CPU)
To run the sample, you need to specify a model: To run the sample, you need to specify a model:

2
docs/articles_en/learn_openvino/openvino_samples/python_sample_throughput_benchmark.md
View File

@ -72,7 +72,7 @@ Running
.. code-block:: sh .. code-block:: sh
python throughput_benchmark.py <path_to_model> python throughput_benchmark.py <path_to_model> <device_name>(default: CPU)
To run the sample, you need to specify a model: To run the sample, you need to specify a model:

12
samples/cpp/benchmark/throughput_benchmark/main.cpp
View File

@ -22,8 +22,12 @@ int main(int argc, char* argv[]) {
try { try {
slog::info << "OpenVINO:" << slog::endl; slog::info << "OpenVINO:" << slog::endl;
slog::info << ov::get_openvino_version(); slog::info << ov::get_openvino_version();
if (argc != 2) {
slog::info << "Usage : " << argv[0] << " <path_to_model>" << slog::endl; std::string device_name = "CPU";
if (argc == 3) {
device_name = argv[2];
} else if (argc != 2) {
slog::info << "Usage : " << argv[0] << " <path_to_model> <device_name>(default: CPU)" << slog::endl;
return EXIT_FAILURE; return EXIT_FAILURE;
} }
// Optimize for throughput. Best throughput can be reached by // Optimize for throughput. Best throughput can be reached by
@ -31,10 +35,10 @@ int main(int argc, char* argv[]) {
ov::AnyMap tput{{ov::hint::performance_mode.name(), ov::hint::PerformanceMode::THROUGHPUT}}; ov::AnyMap tput{{ov::hint::performance_mode.name(), ov::hint::PerformanceMode::THROUGHPUT}};
// Create ov::Core and use it to compile a model. // Create ov::Core and use it to compile a model.
// Pick a device by replacing CPU, for example MULTI:CPU(4),GPU(8). // Select the device by providing the name as the second parameter to CLI.
// It is possible to set CUMULATIVE_THROUGHPUT as ov::hint::PerformanceMode for AUTO device // It is possible to set CUMULATIVE_THROUGHPUT as ov::hint::PerformanceMode for AUTO device
ov::Core core; ov::Core core;
ov::CompiledModel compiled_model = core.compile_model(argv[1], "CPU", tput); ov::CompiledModel compiled_model = core.compile_model(argv[1], device_name, tput);
// Create optimal number of ov::InferRequest instances // Create optimal number of ov::InferRequest instances
uint32_t nireq = compiled_model.get_property(ov::optimal_number_of_infer_requests); uint32_t nireq = compiled_model.get_property(ov::optimal_number_of_infer_requests);
std::vector<ov::InferRequest> ireqs(nireq); std::vector<ov::InferRequest> ireqs(nireq);

11
samples/python/benchmark/throughput_benchmark/throughput_benchmark.py
View File

@ -30,18 +30,21 @@ def main():
log.basicConfig(format='[ %(levelname)s ] %(message)s', level=log.INFO, stream=sys.stdout) log.basicConfig(format='[ %(levelname)s ] %(message)s', level=log.INFO, stream=sys.stdout)
log.info('OpenVINO:') log.info('OpenVINO:')
log.info(f"{'Build ':.<39} {get_version()}") log.info(f"{'Build ':.<39} {get_version()}")
if len(sys.argv) != 2: device_name = 'CPU'
log.info(f'Usage: {sys.argv[0]} <path_to_model>') if len(sys.argv) == 3:
device_name = sys.argv[2]
elif len(sys.argv) != 2:
log.info(f'Usage: {sys.argv[0]} <path_to_model> <device_name>(default: CPU)')
return 1 return 1
# Optimize for throughput. Best throughput can be reached by # Optimize for throughput. Best throughput can be reached by
# running multiple openvino.runtime.InferRequest instances asyncronously # running multiple openvino.runtime.InferRequest instances asyncronously
tput = {'PERFORMANCE_HINT': 'THROUGHPUT'} tput = {'PERFORMANCE_HINT': 'THROUGHPUT'}
# Create Core and use it to compile a model. # Create Core and use it to compile a model.
# Pick a device by replacing CPU, for example MULTI:CPU(4),GPU(8). # Select the device by providing the name as the second parameter to CLI.
# It is possible to set CUMULATIVE_THROUGHPUT as PERFORMANCE_HINT for AUTO device # It is possible to set CUMULATIVE_THROUGHPUT as PERFORMANCE_HINT for AUTO device
core = ov.Core() core = ov.Core()
compiled_model = core.compile_model(sys.argv[1], 'CPU', tput) compiled_model = core.compile_model(sys.argv[1], device_name, tput)
# AsyncInferQueue creates optimal number of InferRequest instances # AsyncInferQueue creates optimal number of InferRequest instances
ireqs = ov.AsyncInferQueue(compiled_model) ireqs = ov.AsyncInferQueue(compiled_model)
# Fill input data for ireqs # Fill input data for ireqs