Vyacheslav
Slate, UAD, Eventide, Sonalksis, Ample, Vienna Ensemble Pro 5, Steven Slate Drums 4, Kontakt, MusicLab RealGuitar, VirtualGuitarist 2, Sonitus, AOM...

Load balancing doesn't apply to your instruments so if the bulk of your load was coming from them then it may not significantly help. What does your CPU load before and after load balancing look like? What is your audio buffer size in preferences?

Noel Borthwick [Cakewalk]

Vyacheslav
Slate, UAD, Eventide, Sonalksis, Ample, Vienna Ensemble Pro 5, Steven Slate Drums 4, Kontakt, MusicLab RealGuitar, VirtualGuitarist 2, Sonitus, AOM...

Load balancing doesn't apply to your instruments so if the bulk of your load was coming from them then it may not significantly help.
1. What does your CPU load before and after load balancing look like?
2. What is your audio buffer size in preferences?

Thanks, Noel!
1. Before - It looks like without this update: 1st kernel (core?) is loaded, the other not.
2. 256.

dwardzala
Perhaps your performance issues aren't due to CPU usage?

I do not think. The configuration of the studio in the signature.

Vyacheslav
1. If I do not work with balanced plug-ins (which obviously), CPU indicator should be showing an uneven load on the eggs .., sorry, the cores!
2. And If the CPU display shows a balanced intake, then he's lying!
Is not it so?

not correct. my cores used to be very balanced in many projects even before load balancing - because Sonar had multi-processor support before and when using similar plugins across many tracks, it did already do a good job on distributing the load. the new load balance should be applied if you have a few tracks which have a lot of plugs alongside tracks that only use a few plugs. it helps reading the documentation because it is very clearly explained there (e.g. when to enable)

Load balancing is only going to help performance if unbalanced loads were the problem in the first place.
 
You can test this by pulling up a problematic project with load balancing disabled. If your CPU is running out of steam, and you see that one core is pegged doing all the work, then better load balancing may solve the problem by reducing the load on that overworked first core. Assuming, as Noel notes above, that the plugins placing the heaviest load on the CPU are amenable to load balancing.
 
Bear in mind that not all glitches are caused by CPU exhaustion. If you find that no cores are maxed out regardless of the load-balancing option and there are still dropouts, that may indicate that the real problem is with a hardware interrupt handler.

in previous versions of Sonar, an FX bin would sometimes use all of the same Core.
with load balancing, FX bins with multiple plug ins, each plug in is spread over the multicore vs. using only the first core, then over flowing to the next.
I hope I explained this correctly, what im trying to say is..if you only have 1 FX per bin, you wont see a dramatic effect on Core usage.

my cores used to be very balanced in many projects even before load balancing - because Sonar had multi-processor support before and when using similar plugins across many tracks

 sonar229030
 sonar229036
It is one and the same project.
 

if you only have 1 FX per bin, you wont see a dramatic effect on Core usage.

If there's no change in performance when you adjust everything else, maybe your issues could be hard drive related? I can't see it very clearly on my phone, but you hard drive meter seems to be getting up there. Just a tought.

Bear in mind that not all glitches are caused by CPU exhaustion. If you find that no cores are maxed out regardless of the load-balancing option and there are still dropouts, that may indicate that the real problem is with a hardware interrupt handler.

Thank you, bitflipper. I understand it.

pinguinotuerto
If there's no change in performance when you adjust everything else, maybe your issues could be hard drive related? I can't see it very clearly on my phone, but you hard drive meter seems to be getting up there. Just a tought.

Thank you, pinguinotuerto. But the same glitches occurred with less filled disc.

Thanks for the animated screenshot, Vyacheslav. A picture is worth a thousand words.
 
As you can see, load balancing really did spread the load across the cores, and the first core isn't nearly as busy in the "after" picture. However, even in the "before" picture, the first core is nowhere near maxing out.
 
A CPU is either capable of handling the current demands or it isn't - it's not like a human, who benefits from a lighter workload even if he isn't being pushed to the limit. As long as the CPU can keep up, you'll experience no subjective difference between the before and after scenarios, which is exactly what you noted.