i’ve just made a quick video and uploaded it to youtube, showcasing the change in CPU load when adding or removing cores from/to renoise’s audio engine on a quadcore.
as you can see, there definately is a great difference in performance between the different amount of utilized cores (in this ideal scenario at least).
Just for shits n grins… I decided to do a little test on my 16 core rig…
System has 4 x opteron 8347 (1.9ghz), 16 gb ram
and after some rummaging around in boxes, a soundblaster live 5.1 digital
and after some time spent playing “find that driver” I killed the end of level bad guy and had a working soundcard under server 2k8…
So I set up 16 tracks, each with its own vst, for testing purposes used Korg Polysix (I think thats what someone else used), and each playing a sustaining bass sound (01Thickbass)…
Just a single pattern, each one playing C4
Then it was time to play…
ENABLE THAT CPU CORE !
Here were my testing results…
As you can see. the percentage would jump about a bit… I simply noted the lowest I could spot, and the highest. I dont have anything else running on the machine bar notepad.
Quite interesting results,. there seems to be a definite sweet spot at 9 cores. probably 2 vst’s per core ?
What was surprising was the increase in percentage as I went from 10 to 16.