Our main operator computers will often have 4 vision clients open at once. Although I think the speed is fine, I get complaints that it is not as fast as our previous wonderware hmi. I've tried running multiple clients on my workstation and it IS faster in general. The only difference between my computer and theirs is that I have a dedicated GPU. Does vision even make use of the GPU? If so, is there documentation showing the requirements?
It is also worth noting that the operators use 3 monitors. Two smaller 24in and one 43in. These monitors are display port daisy chained together. I know there are PLENTY of different things that can affect Ignition's performance, but I wanted to start at the hardware and see if I am missing something obvious.
Vision uses Java Swing, which does generally speaking use hardware acceleration for graphics but generally speaking hardware isn't the most common bottleneck for Vision performance unless you're on seriously underperforming hardware.
Typically operators reporting bad performance in Vision are actually experiencing errors in project design leading to things locking up the main GUI thread, which manifests as poor response times. Is there lots of scripting used in your project?
There is a fair amount of scripting. I have actually changed a bunch of our scripts over the past year to try and increase performance. I suppose it would be a good idea to get in touch with support and see if they could look over my scripts.
In general I've been trying to increase performance in every way I can. I had never considered if the hardware could be bogging things down but I doubt that is the case. Thanks for the response.
I just thought of something else. The operator computers that I spoke of above all run a VM (locally) with Hyper-V for some legacy software that we have that doesn't run on windows 10. The VMs themselves are not too taxing (2 vCPUs and 4GB RAM). These computers should be beefy enough to deal with it, but I was wondering if you had any opinions on whether running a VM on the same machine could slow down a vision client?
It's certainly possible. There's only so much a physical CPU can do at a time; and you're more likely to notice CPU constraints in 99% of Vision projects than GPU constraints. Might be good to keep an eye on Task Manager/per core CPU usage from the host system, especially while operators are reporting bad performance.