I have a question about predictive loading. For years I had it turned off, but my gig file grew so big that I was starting to hit the ceiling of my memory limits, I decided to try predictive loading. And indeed, when I first got started, it not only opens faster, but it uses a lot less memory. However, it seems to compile as I switch rackspaces. I had assumed that since it only loads those rack spaces that are within limit set in the settings. For example, if it was set to one, then when you went to another one, it would kind of unload the first one and load the second one and relatively speaking, the memory usage would stay about the same (of course depending on the given rack space but in my case they are relatively similar). But that’s not what is happening. As I keep switching to new rackspaces or songs, the memory usage compiles and grows higher and higher. As I watch the memory usage in task manager, it may start out under 3 GB of RAM and then it works its way up to 10 GB eventually. Is this the way it’s supposed to be or I got something set wrong somewhere?
I think that depends on if a plugin frees up memory when released.
I did put this new plug in I really like all over the place, I’m wondering if that’s why my memory issue started to happen lately? I’ll test that by removing it. Could it also be memory going bad or bad Windows memory management?
Perhaps you might share the name of this new plugin ?
It’s NAM Universal 2. It’s one of the best Neural Amp Model loaders, but it is still in beta. It’s been solid for me so I used it in a lot of RS but I didn’t think about it potentially not releasing memory.
I see this marked as “solution” - does this mean definitively that you took out NAM Universal 2 and the memory buildup stopped?
No, it’s not quite as black and white as that. Nam Universal 2 is one of them but it’s not the only plug-in that does this.