Download Performance-testing Tools for HLA 1516Pitch offers a performance testing tool called “RTIperf 1516” for HLA 1516. You are welcome to download this program and its documentation from this page. No registration is required. This program is not intended to be used with pRTI 1516 LE since it will usually exceed the LE limits (number of updates etc).
Overview of RTIperf 1516This program consists of two applications: a master and a slave. It measures update rate, throughput, latency and loss rate and offers reliable and best-effort communications. The send rate for best effort can be controlled. The program can be run in several modes: As a command line tool where you give commands like
“latency 1000 500” to get the latency for 1000 packets of 500 bytes of data. Using test suites. These are script files consisting of several commands. The output is stored in text files that can be opened in Excel. Standardised test suites are useful if you want to compare different setups in a consistent way. With the graphical user interface. You may then interactively select tests similar to the command line. The chosen test parameters and the results are continuously displayed in a log window. The “remote slave” mode. This enables you to run a slave federate for your tests on a machine that may be remote, for example in another country. If you restart the RTI on your local machine the slave will reconnect within 60 seconds, ready to accept new test runs. RTIperf is written in Java and requires a compliant HLA 1516 RTI (DoD interpretations 2) with Java bindings. The RTI should not have limits in the number of updates/interactions that can be sent.
LicensingRTIperf 1516 is available together with the source code under the GNU General Public License, which means that: You can use the program without charge You get access to the source code You can use the source code to develop new and extended versions of the program You may distribute your extended version of the program You may not sell any version of the program as a commercial product
By downloading the software you accept the license terms. For full information about the license read the full General Public License or visit the GPL web site.
DownloadTo get to know RTIperf 1516 you may want to start by reading the documentation.
Download RTIperf 1516 documentation (PDF, 245 kB)
The installation kit consists of the compiled program, FOM, source code, documentation and a sample test suite. This installation kit is for Windows but since RTIperf is implemented in Java it can easily be run on other operating systems. The default installation is made in a new directory in the standard pRTI 1516 "samples" directory. For other platforms or RTIs you may want to modify this.
Download the RTIperf 1516 installation kit for Windows (500 kB)
SupportNo free support is offered for this program. You are welcome to report bugs, suggest improvements and report about your own versions to
This e-mail address is being protected from spambots. You need JavaScript enabled to view it
.
Here are some helpful hints:
Start by determining what performance your federation needs and how important other factors like reliability, fault-tolerance, debugging facilities etc are compared to performance. Be sure to read the RTIperf documentation carefully before running the program. Run the master and slave on different computers. Know your network so you know what optimal throughput and latency to expect. Check maximum file transfer speeds. Use the "ping" command to find out typical latency values. Run a "warm up" before running your real tests. Enter all necessary parameters (including timeout for best effort) correctly before running tests from the GUI. The program does not do extensive checking for typing mistakes. Run several test runs and compare the result. You should expect some fluctuations. For small data sizes the limitations may be the CPU load and the maximum number of packets the OS can send. Try running with hosts with different CPU speeds. For larger data sizes the limitation may be the network throughput. Try running with different network speeds (10/100/1000 Mbit/s). You will usually not get more than 90 percent of the theoretical capacity on 10/100 Mbit networks. When testing with best effort you should limit the data rate. Otherwise the sender may send data too fast and the loss rate will be high. Also check that you have set the timeout correctly. Be careful with the parameters that you enter. If the total amount of data (the product of the number of updates and the update size) is big the test will take a long time to complete. If you run over WANs (like the Internet) where you share the capacity with others, try running the tests at different times.
|
