Because there is nothing worse than having your system freezing or crashing in the middle of a performance or installation it is important to test its limits. The following guide aims at providing some suggestions to test Puredyne in order to both know how far your machine/software can be pushed, but also to help the project to spot abnormal behaviours.
Stress testing is more meaningful when you use a software that you already know so you have a common ground to compare things. If the software is not present we need to know that it's missing and that you need it, specially if it's a software that used to be in a previous version of Puredyne but that did not make it to a new version.
Similarly we need to know if the system is ready to be used for your testing or if something still needs to be fixed. For example:
- Can you install Puredyne?
- Is the live persistence working?
- Can you boot at all?
- Do you have sound? 3D acceleration?
- Can you get online?
- etc (suggestions welcome)
Focused stress testing
Once you have an environment that is providing the minimum standard for your testing, just use the software you are used to (your fav Pd patch, ardour session, sc code, video editing suite, inkscape orgy, etc) and check if it's working as expected:
- Is it slower than usual?
- Is it crashing more than it should?
- Can you run several instances of it and still have a relatively responsive system (if relevant to the application)?
- Can you work or use the software for a long time or it freezing after a minute, hour, day?
- etc (suggestions welcome)
Try to do something unusual to see how far your system can handle it, for example set Jack to a rather low latency and start piling up jack-rack effects in a complex feedback loop. Open Several applications at the same time, go online and open a hundred tabs in Firefox or Chromium, watch an HD movie, etc.
The point is not to prove that Puredyne will choke. It will choke. But is the choking point acceptable for you or seem normal or seem to be different (for the worse) from another system?
General audio latency test
FIXME - What would be a good protocol to test audio latency on Puredyne and keep a database of results so that everyone can compare with each others from one version to another?
Adopting an alpha/beta system as main
As stress testing relies on a mix of very specific subjective and objective elements, issues not covered by the tests will obviously go under the radar. This is why adopting a work in progress system can be very helpful in spotting missing software, new bugs, etc.
Contribute to this document
Just like the rest of the Puredyne documentation, feel free to edit and modify this page, we would love to hear about specific ways to test Puredyne that are relevant to media practices.Last modified on 12 April 2012, at 22:09