It's still a very simple patch, I'm pretty sure, I#ve made mistakes, 1,6 Ghz with 2 GB of ram must be more than enough for something like this, it's just a little sample player and I'd like to make it a big one. Problem is, sometimes it works for hours, with many channels, many samples and even vsts included, cpu always around 10%, and then, suddenly it climbs up to 100 and stays there, sometimes even if only one channel with a single soundfile is used. I can't refer this to a specific cause, cpu load increases at varoius times, sometimes even when the files are just played back without tweaking. I tried this without pitch envelopes, without freeverbs, with lots of different soundfiles... all 24 bit mono files, 1, 2 or 4 seconds... I ask you to have a quick look at this, maybe play around with it for a little while.... I hope I made a mistake here.....
-
My CPU-Fucker
-
Attached the parent again, something was hidden
-
Just took a quick look. My guess is that it's happening when you stop sending anything to one of the reverb units. Try constantly sending 1 dB of [noise~] into [freeverb~]. It has some denormal issues when it has to crunch numbers that are extremely small, which causes it to eat up cpu. The noise will keep the numbers high enough.
-
Maelstorm, I think I owe you a drink......
-
I'm drunk enough already...but I will accept.
I've recently been put in the situation where I can't really get around programming an external, and I've discovered that Miller implemented a simple macro to prevent denormal issues. If I find the time these next few weeks, I'm going to go through [freeverb~] and submit a fix. It seems long over due.
