Basically you just open the software, set your two conditions at which the software purges your standby list, and let it run in the background.
You have your total memory, standby list & system working set, and free memory. When your game is running, it will use a certain amount of memory (my total is 32.7MB, my game uses around 14k MB. leaving my free memory at 18k MB). I noticed when the game crashes, my standby list jumped to over 8k MB, so I have the software set to purge the standby list at 1MB when my free memory is lower than 18.5k MB.
That is a confusing way of saying that I have the program set to purge my standby list at about 2k MB over what the game uses out of my total memory. Never letting the standby list get abnormally high, and has so far (about 9 hours of gameplay) resulted in me crashing ZERO times…which is astounding (just look at my post history detailing my separate crashes).
edit - I’m aware this may not work for everyone, but it works for me so far on this game and has worked in a previous game years ago having the same issues. Obviously, try at your own risk, but there is little to none for a significant benefit.
Understanding what the ‘standby’ memory list actually is and what is causing it is much more meaning full.
ISLC along with other apps can ‘force’ the ‘standby’ memory to be released, but that doesn’t prevent application from ‘pre-loading’ it back in to memory again.
And you end up with a background load to memory tug-of-war and release standby memory actions happening while your playing a game. Your better off reducing it’s foot print.
Windows prefetch service, ‘Sysmain’ if I recall correctly is part of why your getting a large chunk of data in the ‘standby’ memory list. Any application running, will utilize the standby memory pool and this includes Darktide when it’s running and when it closes you end up with that spike.
It might provide some stability, but if disabling sysmain/prefetch and unloading unneeded application doesn’t work for you, your next best bet is upgrading and adding memory.
I tried this now with a 1 hour session and it did not stop the GPU hang error. The working set never exceeded 11,500 MB (32684 MB total system) during that time and no purge occurred. It was worth a shot.
edit: Will try a manual purge after loading a new map…
edit2: Still occasional crashing on map load, GPU hang error. Can occur with less than 1000 MB in the working set.