Cubendo Daw bench question???

So I just put together a new DAW and wanted to test it out with DAWbench.

The question is are you supposed to turn the Comps on one per track and then return to the first track and add another or add them all on 1 track and go to the next? It gives VERY different results depending…

Sorry if this is the wrong forum but I had to post somewhere :slight_smile:

IT depends on what you want to measure.

Turning on track measures how many of those sine waves you can serve from disk before you have problems. Obviously, there is CPU load from all the tracks.

Turning on all plugs before going to the next track focuses on plug performance as the majority of the load will be taken up by the compressor and not the track/media/disk

OK. Sorry, I might not have explained it properly:

On the Daw bench site it measures the performance of the new i7 and other processors by how many Comps you can turn on before getting pops or glitches.

The issue is if you turn on all the comps on one track it puts WAY more load on the processor than just turning on one each track.

Thare’s not alot of info on the site is all…

I didn’t get that behavior. In fact, I got the opposite.

8 comps on 1 track vs 1 comp on 8 tracks was virtually identical.

It was only when I got up to 50 or 60 tracks that the track load started overtaking the comp load in a noticeable way.

Are you sure you are using real copies of the loop on each track and not just reference copy? If you copied to create more tracks, you need to convert to real.

The readme text says you should turn on one MBC per track until you’ve reached all 40 tracks, then go back and turn a second MBC on each track. They explained that it spreads the load out more evenly. This is right out of the readme file…

“While the session is playing, progressively turn on the first insert plugin on each channel- going horizontally across from channel 1-40, this will distribute the load more evenly. Note: The method of activating the plugins varies between applications, check the correct procedure for the respective app. If/When you reach the 40th Sine channel , scroll back to the first Sine channel and activate the second insert plugin, and then move on to the next Channel and do the same, etc. Continue until you clearly hear audio breakup. At that point solo the Monitor Channel, raise the fader double check for any breakup. Remove some plugins until no audio breakup is heard, and then reactivate the plugins again if possible.”

right, that’s so that the streaming load is not noticeable within the overall # of plugins.

Look at it this way. You will get a lesser compressor count if you kept adding 1 track with 1 compressor than if you put 2 compressors per track, which will be less than you would get if you put 3 per track etc…

There is a point where you system can’t stream the track data regardless of the # of plugs. This will be variable based on RAM and HD configuration. The question is, do you want to measure Plugin Count or Track Load? Or a reasonable cross section. The default way they say to do it is a reasonable cross section.

Yeah… I guess readme files are the thing to read before posting (duh).

Thanks for the responses. I must say that I’m getting great results with my new system at stock speeds.

[i7 950, Gigabyte GA-X58A-UD5 rev2, M-Audio Profire 2626, 24 Gigs Kingston Hyper X ram]