[isaacsim.benchmark.services] Benchmark Services#
Version: 2.6.3
This extension provides benchmarking utilities
Enable Extension#
The extension can be enabled (if not already) in one of the following ways:
Define the next entry as an application argument from a terminal.
APP_SCRIPT.(sh|bat) --enable isaacsim.benchmark.services
Define the next entry under [dependencies]
in an experience (.kit
) file or an extension configuration (extension.toml
) file.
[dependencies]
"isaacsim.benchmark.services" = {}
Open the Window > Extensions menu in a running application instance and search for isaacsim.benchmark.services
.
Then, toggle the enable control button if it is not already active.
API#
Python API#
Front-end for benchmarking standalone scripts and other non-async snippets. |
- class BaseIsaacBenchmarkAsync(*args: Any, **kwargs: Any)#
Bases:
AsyncTestCase
- async setUp(backend_type: str = 'JSONFileMetrics')#
Must be awaited by derived benchmarks to properly set up the benchmark
- async tearDown()#
Must be awaited by derived benchmarks to properly tear down the benchmark
- set_phase(
- phase: str,
- start_recording_frametime: bool = True,
- start_recording_runtime: bool = True,
Sets benchmarking phase. Turns on frametime and runtime collection.
- Parameters:
phase (str) – Name of phase, used in output.
start_recording_frametime (bool) – False to not start recording frametime at start of phase. Default True.
start_recording_runtime (bool) – False to not start recording runtime at start of phase. Default True.
- async store_measurements(
- stop_recording_time: bool = True,
Stores measurements, metadata, and artifacts collected by all recorders during the previous phase. Optionally, ends frametime and runtime collection.
- Parameters:
stop_recording_time (bool) – False to not stop recording runtime and frametime at end of phase. Default True.
- async fully_load_stage(usd_path)#
Await this function to open a stage and then wait for it to fully load
- Parameters:
usd_path (str) – Path to USD stage.
- async store_custom_measurement(
- phase_name: str,
- custom_measurement: <module 'isaacsim.benchmark.services.metrics.measurements' from '/builds/omniverse/isaac/omni_isaac_sim/_build/linux-x86_64/release/exts/isaacsim.benchmark.services/isaacsim/benchmark/services/metrics/measurements.py'>,
Stores the custom measurement specificed in the benchmark.
- Parameters:
phase (str) – The phase name to which the measurement belongs.
measurement (Measurement) – The measurement object to store.
- class BaseIsaacBenchmark(
- benchmark_name: str = 'BaseIsaacBenchmark',
- backend_type: str = 'OmniPerfKPIFile',
- report_generation: bool = True,
- workflow_metadata: dict = {},
Bases:
object
Front-end for benchmarking standalone scripts and other non-async snippets.
By default this class will collect hardware performance data (see recorders), broken up by “phase”. Typical structure looks like:
benchmark = BaseIsaacBenchmark(benchmark_name=..., workflow_metadata=...) benchmark.set_phase("loading") # load stage, configure sim, etc. benchmark.store_measurements() benchmark.set_phase("benchmark") # Actual code being benchmarked (running the sim for N frames, cloning an object, etc.) benchmark.store_measurements() benchmark.stop() # Shuts down benchmark, writes metrics to file
You can set any number of phases.
- stop()#
Stop benchmarking and write accumulated metrics to file.
- set_phase(
- phase: str,
- start_recording_frametime: bool = True,
- start_recording_runtime: bool = True,
Sets benchmarking phase. Turns on frametime and runtime collection.
- Parameters:
phase (str) – Name of phase, used in output.
start_recording_frametime (bool) – False to not start recording frametime at start of phase. Default True.
start_recording_runtime (bool) – False to not start recording runtime at start of phase. Default True.
- store_measurements(
- stop_recording_time: bool = True,
Stores measurements, metadata, and artifacts collected by all recorders during the previous phase. Optionally, ends frametime and runtime collection.
- Parameters:
stop_recording_time (bool) – False to not stop recording runtime and frametime at end of phase. Default True.
- fully_load_stage(usd_path: str) None #
Loads provided USD stage, blocking execution until it is fully loaded.
- Parameters:
usd_path (str) – Path to USD stage.
- store_custom_measurement(
- phase_name: str,
- custom_measurement: <module 'isaacsim.benchmark.services.metrics.measurements' from '/builds/omniverse/isaac/omni_isaac_sim/_build/linux-x86_64/release/exts/isaacsim.benchmark.services/isaacsim/benchmark/services/metrics/measurements.py'>,
Stores the custom measurement specificed in the benchmark.
- Parameters:
phase (str) – The phase name to which the measurement belongs.
measurement (Measurement) – The measurement object to store.
Settings#
Extension Settings#
The table list the extension-specific settings.
Setting name |
Description |
Type |
Default value |
---|---|---|---|
|
Default test suite name. |
|
|
|
Whether to add a randomly generated string as a prefix to the output filename to distinguish runs. |
|
|
The extension-specific settings can be either specified (set) or retrieved (get) in one of the following ways:
Define the key and value of the setting as an application argument from a terminal.
APP_SCRIPT.(sh|bat) --/exts/isaacsim.benchmark.services/SETTING_NAME=SETTING_VALUE
Define the key and value of the setting under [settings]
in an experience (.kit
) file or an extension configuration (extension.toml
) file.
[settings]
exts."isaacsim.benchmark.services".SETTING_NAME = SETTING_VALUE
Define the key and value of the setting using the carb framework (in Python).
import carb
settings = carb.settings.get_settings()
settings.set("/exts/isaacsim.benchmark.services/SETTING_NAME", SETTING_VALUE)
Define the key to query the value of the setting using the carb framework (in Python).
import carb
settings = carb.settings.get_settings()
value = settings.get("/exts/isaacsim.benchmark.services/SETTING_NAME")