npx skills add ...
npx skills add nvidia/omniperf --skill benchmark-isaacsim
Run Isaac Sim benchmark scripts and interpret benchmark outputs. Covers camera, SDG, scene-loading, robot, lidar/radar/sensor benchmark scripts, common parameters, output files, and benchmark-specific pitfalls. Use when the user asks to run or compare Isaac Sim benchmark results. NOT for initial bottleneck triage (use diagnose-perf), profiling capture (use profiling), trace analysis (use nsys-analyze), or applying performance fixes (use perf-tuning).
npx skills add nvidia/omniperf --skill benchmark-isaacsim
Parameter references may be outdated. Always verify with
./python.sh <script> --help. For profiling details (Tracy, Nsight), see theprofilingskill.
See the install-isaacsim skill for installation (pip, source build, Docker).
profiling skill.perf-tuning skill.install-isaacsim skill.profiling skill.All in standalone_examples/benchmarks/. Run via ./python.sh <script> [args].
| Script | What it measures | Key params |
|---|---|---|
benchmark_camera.py | Multi-camera rendering FPS | --num-cameras, --resolution W H |
benchmark_sdg.py | Synthetic data generation throughput | --num-cameras, --annotators, --asset-count |
benchmark_scene_loading.py | Scene load time + FPS | --env-url (required) |
benchmark_robots_o3dyn.py | O3Dyn robot physics | --num-robots, --physics |
benchmark_robots_humanoid.py | Humanoid physics | --num-robots, --physics |
benchmark_robots_nova_carter.py | Nova Carter (no ROS) | --num-robots |
benchmark_robots_nova_carter_ros2.py | Nova Carter + ROS2 | --num-robots, --enable-3d-lidar, --enable-hawks |
benchmark_robots_evobot.py | Evobot multi-phase | --num-robots 1 10 20 |
benchmark_robots_ur10.py | UR10 manipulation | --num-robots, --device |
benchmark_core_world.py | Core world cloning | --num-envs (no --num-frames) |
benchmark_rtx_lidar.py | RTX lidar sensor | --num-sensors, --lidar-type |
benchmark_physx_lidar.py | PhysX lidar sensor | --num-sensors |
benchmark_rtx_radar.py | RTX radar sensor | --num-sensors |
benchmark_single_view_depth_sensor.py | Single-view depth camera | --num-cameras, --resolution W H |
benchmark_rtx_lidar_ros2_pcl_metadata.py | RTX lidar + ROS2 PCL (v6+) | --num-sensors, --metadata |
benchmark_nucleus_kpis.py | Nucleus KPIs | (none) |
Common params: --num-frames (default 600), --num-gpus, --backend-type, --viewport-updates, --non-headless
Output control Kit args (append to any benchmark command):
Isaac Sim uses hyphens in parameter names. Underscores are silently ignored by argparse:
Always verify result JSON to confirm params were applied.
When comparing benchmark outputs, record whether the run used --non-headless, --viewport-updates, render products, or camera sensors. Do not change viewport code or settings from this skill; if extra viewport work appears to affect results, hand off to perf-tuning.
Expired OMNI_PASS tokens cause silent asset loading failure. See install-isaacsim skill for the expiry check command.
Monitor progress by polling for result JSON files, not watching stdout.
Kit log: ~/.nvidia-omniverse/logs/Kit/isaacsim*/kit.log or --/log/file=/tmp/kit.log.
If kpis_*.json and *.tracy exist with non-zero size and the process has not exited after 2 minutes with no new output, follow the profiling skill's guide-aligned shutdown handling and force-kill the app and capture processes.
If results indicate per-camera viewport overhead, too many render products, default viewport work in headless mode, or poor multi-GPU scaling, do not apply fixes here. Summarize the benchmark evidence and use perf-tuning for the configuration changes.
kpis_benchmark_<name>.json — benchmark KPIs (main results)kit.log — execution log (if --/log/file= is set)*.tracy — Tracy profiling trace (only with Tracy args)*.nsys-rep — Nsight profiling trace (only with nsys)Note:
benchmark_result.jsonandstartup_*.jsonare NOT produced by default in v5.x/v6.x. The primary results file iskpis_benchmark_<name>.json.