.NET Application very slow after long period of inactivity

Question

I'm not sure if the question title is the best one but it was the best one I could come up with...

I have this .NET (C#) applications which starts up with Windows and remains opened until the computer is turned off. The app stays on the tray and I open it by clicking the tray icon and close it the same way.

The app is not slow at first, it works normally, no problems there. But after long periods of inactivity of itself, it gets very slow when showing it again for the first time in a long period. Know what I mean.

For instance, I could not use/open (click the tray icon) for a few days and between those days I opened and closed and used lots of other apps, heavy apps too and I probably hibernated and resumed the computer a few times and when I needed to open my app again, it was slow. After a few minutes of using it, it goes back to normal and works fine.

I believe this has something to with memory management and the system probably frees up most of my app's memory so other programs can use it more efficiently. And maybe .NET memory management as something to do with it...

Whatever the reason, is there anything I can do to optimize my app regarding that issue?

Answer 1

This is almost certainly due to memory being paged out to disk.

When you cease using your application and other applications or tasks start exerting memory pressure, pages of memory from your app can be written out to disk. When you try to use your application again, all this data must then be read in causing the stalls that you see.

Unfortunately, there is no good solution - if you run as administrator or have the SeLockMemoryPrivilege, you can lock parts of your application into physical memory. You can try "touching" pages periodically to keep them in physical memory. Unfortunately, both these options will cause your application to interact badly with other applications on the system - your memory is getting paged out because the physical memory is needed for something else. You can attempt to lower your overall memory footprint, but you will still run into this issue in some cases. There are options for tweaking your working set size, but the OS is free to ignore those, and will in many cases.