Subject says it all, but I'm wondering what would give the best performance for a large orchestral template: running many different instances of VEP (say 6-10 instances) or one very large instance. Would it make a difference? If multi-instane is the way to go, how would you configure the multi-threading for each instance?
Thanks!