Hi, I have been looking for a solution to this problem for a long time. basically I have an 8 core processor, but most vst hosts or visual programming environments I know of use only one processor core. It seems parallel processing of a changeable graph is a pretty hard scheduling problem. How does gig performer manage multiple cores?
Gig Performer is not a visual programming environment. The “graph” is simply a more convenient way to specify the flow of audio through multiple plugins.
When a plugin is instantiated, it typically creates its own thread(s) to do its work. One of the roles of the OS is to distribute threads across available cores. What exactly are you trying to do that is causing Gig Performer to not work because of this concern?
Hi David, and again thanks for the response. So far gig performer has worked fine, but I have only made a single rack with a single plugin. I have tried Max, Bidule, and Hollyhock Usine for live setups. Bidule performs the best, but has difficulty with multiple cores because the graph can change on the fly, so unless you tell it otherwise, all the processing threads remain on one core. Maybe the rack model sufficiently isolates things that the racks can be scheduled on different cores. I am not a programmer, so I am not sure I have the terms right, but the best I understand it, multiple threads get scheduled in the other three programs on one core until told to do otherwise, and when assigned to multiple cores they may or may not work. I was just wondering how Gig performer allocates threads across the cores before I invest a lot of time in making complex setups (which is a habit of mine!). I hope that is clear, but this is a query, not a bug.
Richard, may I suggest that you try creating a few rackspaces and see how they work. Only the plugins in the active rackspace use CPU cycles so when you switch to another rackspace, all the plugins in the first rack STOP.