Plugin Example is the reference RoboDK Plug-in used as a starting point for developing new plug-ins. It demonstrates the
Plug-in Interface: registering menu/toolbar actions, adding a docked window, listening to RoboDK events, and
calling the RoboDK API (IRoboDK/IItem) directly from C++ for maximum performance.
- Plugin Speed Information (
Ctrl+I): runs a benchmark of the RoboDK API on the selected robot (Forward Kinematics, Inverse Kinematics, and collision checking) and reports the results in a docked window. - Program collision check: optionally checks every step of a program for collisions and reports how many points are in collision vs. collision-free, along with timing statistics.
- Includes a System / CPU / RAM summary of the computer running the benchmark.
- Robot Pilot Form: a docked window to jog the robot by incremental steps (joints or Cartesian, relative to the tool or the reference frame).
- Right-click integration: the benchmark and robot pilot actions are added to the context menu of objects and robots in the station tree.
- Select Tools → Plug-Ins in RoboDK and double click on PluginExample if it is not already loaded.
- Use the Plugin Example menu or toolbar:
- Plugin Speed Information (or
Ctrl+I): pick a robot to benchmark. You can then optionally pick a program to run a collision check against; if a program namedMainexists, it is used automatically. - Robot Pilot Form: opens the jog panel described above.
- RoboDK Plugins - Help: opens the RoboDK Plug-in documentation in your browser.
- Plugin Speed Information (or
- Results are displayed as a table in a docked window and are also printed as an aligned plain-text table in the console/debug output, which is useful when there is no GUI available (see below).
| File | Description |
|---|---|
pluginexample.h / .cpp |
Plugin entry point: IAppRoboDK implementation, menu/toolbar setup, benchmark logic |
formrobotpilot.h / .cpp / .ui |
Robot Pilot docked widget |
run_api_benchmark.py |
Standalone Python script that runs the same kind of benchmarks using the RoboDK API for Python |
run_plugin_benchmark.py |
Python helper that downloads the sample station, starts a headless RoboDK instance, loads this plugin, and runs the benchmark automatically via robolink |
manifest.xml |
Add-in package metadata |
The benchmark can be triggered without any user interaction using RoboDK's -PluginCommand argument, which
calls PluginCommand("BenchmarkInfo", progname). This is useful to collect performance stats headlessly, for
example as part of an automated test or CI (Continuous Integration) workflow:
./RoboDK -NEWINSTANCE -NOUI -SKIPINI -PLUGINLOAD=PluginExample "C:/RoboDK/Library/Welding-with-Comau-Smart5-NJ-130-2-6.rdk" -PluginCommand=BenchmarkInfo=MainProg-NEWINSTANCE: starts a new RoboDK instance instead of reusing one that is already running.-NOUI: runs RoboDK without showing its main window.-SKIPINI: skips loading the user's saved settings/station list.-PLUGINLOAD=PluginExample: loads this plugin on startup.- The quoted path is the RoboDK station (
.rdk) to open. -PluginCommand=BenchmarkInfo=MainProg: runs the benchmark against the program namedMainProgand prints the results as text to the console. Since-NOUImeans the docked report is never shown, this text output is the only way to read the stats in this scenario. To save the output to a file when triggering the action manually, launch RoboDK with the-DEBUGflag (C:/RoboDK/RoboDK-Debug.baton Windows).
Alternatively, run_plugin_benchmark.py automates the same steps via the Python robolink API: it downloads
the sample station, starts a headless RoboDK instance, loads the plugin, and streams the benchmark output to
stdout.
The following results were obtained using the Spot welding station with Comau and RoboDK v6.0.6. RoboDK v6 includes important performance improvements for collision checking compared to previous versions.
Note: These results illustrate RoboDK performance on the systems tested and should not be interpreted as a controlled comparison of processors or operating systems. Performance may vary depending on the RoboDK build, station complexity, collision settings, system configuration, power settings, and background processes.
| System | Forward Kinematics | Inverse Kinematics | Collision rate (5 samples) | Collision rate (full program) |
|---|---|---|---|---|
| i9-14900KF — Windows (Desktop) | 0.61 µs | 2.00 µs | 744 samples/sec | 219 samples/sec |
| i7-1165G7 — Windows (Laptop) | 1.57 µs | 11.36 µs | 524 samples/sec | 83 samples/sec |
| Apple M5 Pro — native build | 0.30 µs | 1.39 µs | 4357 samples/sec | 1204 samples/sec |
| Apple M1 — native build | 0.99 µs | 3.59 µs | 4196 samples/sec | 725 samples/sec |
| Apple M1 — Intel build (Rosetta 2) | 2.09 µs | 6.16 µs | 3219 samples/sec | 707 samples/sec |
Metric Value
--------------------------------------------------
Robot Comau Smart5 NJ 130-2.6
System Windows 11 Version 2009
CPU Intel(R) Core(TM) i9-14900KF, 32 cores @ 3.19 GHz
RAM 31.8 GB
Forward Kinematics 0.61 microseconds
Inverse Kinematics 2.00 microseconds
Inverse Kinematics (all solutions) 2.23 microseconds
Collision check (5 samples) 1.34 ms/sample
Collision check rate 743.55 samples/sec
Points with collisions 0
Points without collisions 5
-- Program Collision Check: main --
Collision check (3589 steps) 4.57 ms/step
Collision check rate 218.99 samples/sec
Points with collisions 66
Points without collisions 3523
Metric Value
--------------------------------------------------
Robot Comau Smart5 NJ 130-2.6
System Windows 11 Version 2009
CPU 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz, 8 cores @ 2.80 GHz
RAM 15.6 GB
Forward Kinematics 1.57 microseconds
Inverse Kinematics 11.36 microseconds
Inverse Kinematics (all solutions) 11.88 microseconds
Collision check (5 samples) 1.91 ms/sample
Collision check rate 523.58 samples/sec
Points with collisions 0
Points without collisions 5
-- Program Collision Check: main --
Collision check (3589 steps) 12.07 ms/step
Collision check rate 82.88 samples/sec
Points with collisions 66
Points without collisions 3523
Metric Value
--------------------------------------------------
Robot Comau Smart5 NJ 130-2.6
System macOS Tahoe (26.5.2)
CPU Apple M5 Pro, 15 cores
RAM 24.0 GB
Forward Kinematics 0.30 microseconds
Inverse Kinematics 1.39 microseconds
Inverse Kinematics (all solutions) 1.36 microseconds
Collision check (5 samples) 0.23 ms/sample
Collision check rate 4357.14 samples/sec
Points with collisions 0
Points without collisions 5
-- Program Collision Check: main --
Collision check (3589 steps) 0.83 ms/step
Collision check rate 1203.55 samples/sec
Points with collisions 66
Points without collisions 3523
Metric Value
--------------------------------------------------
Robot Comau Smart5 NJ 130-2.6
System macOS Tahoe (26.5)
CPU Apple M1, 8 cores
RAM 16.0 GB
Forward Kinematics 0.99 microseconds
Inverse Kinematics 3.59 microseconds
Inverse Kinematics (all solutions) 2.65 microseconds
Collision check (5 samples) 0.24 ms/sample
Collision check rate 4196.39 samples/sec
Points with collisions 0
Points without collisions 5
-- Program Collision Check: main --
Collision check (3589 steps) 1.38 ms/step
Collision check rate 724.92 samples/sec
Points with collisions 66
Points without collisions 3523
Metric Value
--------------------------------------------------
Robot Comau Smart5 NJ 130-2.6
System macOS Tahoe (26.5)
CPU Apple M1, 8 cores @ 2.40 GHz
RAM 16.0 GB
Forward Kinematics 2.09 microseconds
Inverse Kinematics 6.16 microseconds
Inverse Kinematics (all solutions) 5.24 microseconds
Collision check (5 samples) 0.31 ms/sample
Collision check rate 3219.49 samples/sec
Points with collisions 0
Points without collisions 5
-- Program Collision Check: main --
Collision check (3589 steps) 1.41 ms/step
Collision check rate 707.07 samples/sec
Points with collisions 66
Points without collisions 3523