Anyone care to chime in with their experience so far?
With people who are seeing a large reduction in ASIO metering/usage in C8, how many cores are you allotting to Kontakt running inside C8, with ASIO Guard 2 enabled? Anybody done any experiments with different K5 core settings yet with respect to AG2?
After extensive testing with C7.5 here, I determined that, at least on my rigs, we had the highest load capacity and best overall performance by running everything [and I mean, EVERYTHING] in VEP on a local server instance outside of Cubase with C7.5 multithreading off, AG turned off, two threads per V-frame, everything divided into 12 v-frames, and multiprocessor off in K5. This is on a 2012 12-core MP running 10.8.5 with 96 gigs of RAM.
However… with all the ballyhooing about AG2 functioning similarly to Logic’s buffer now [I was on Logic for 20-ish years, starting with Notator, but finally gave it up for C7.5 compared to LPX], it’s looking like I should [ugh] re-think my template yet again, and adopt some of my prior best practices from LP9; namely to run more single mono-timbral instances of everything rather than multis; run most everything local inside the sequencer [I have multiple fully stand-alone slaves for the orchestral heavy-lifting, so I only run things that change from cue to cue in the main comp/sequencer - drums, specialty instruments, sound design, synths, etc. - it’s still a lot]. Steinberg seems to be recommending using more single instances with AG2, as it will calculate a multitimbral inst and track chain all on the record buffer once any midi track is monitor/record-enabled that points to it.
On that note, I wonder if, similar to Logic, everything downstream of a record-enabled inst. also gets calculated on the same thread with this new scheme [the dreaded one-core Logic-style overloads, anyone? ] What if multiple vsti’s then have sends on their tracks to the same reverb/delay, etc.? Do all of them, and the return efx, shift to be processed on the record buffer too?
Alright everyone - please weigh in with your current experience, findings, and settings. Let’s see if we can find some consensus on the best way to maximize things under the newly improved AG2.