I just deleted the Mackie CS
Checkd cpu
And the opened the dialog to Add the ACT

ACT does the same thing.


As for the MIDI light being on, the Mackie CS is the same as S5, there is no new functionality in S6 that would change how the MIDI indicator works

ORIGINAL: xackley

Error in the MIDI Activity.

You are correct it is on solid only when I have the ACT surface configured.
With just the Mackie, it is staying off now.

But there is no change in the 10% cpu usage for the Mackie or ACT CS.

I mentioned the MIDI activity only as a possible clue.

Retract the admission of error on reporting the MIDI activity.
As soon as I openned the Mackie Properties page, the MIDI activity light on the right went back to Solid, and stays that way.

Checked in S5 and it stays off unless a control is moved using the Mackie with equal setting in the property page.

Has any other forum user, besides KEVO, bothered to check If Control Surfaces are using a noticeable amount of CPU in Sonar 6.

yes, i can confirm that i am getting the same behavior with my MCU. note i no longer have sonar 5 installed so i couldnt check that.

Sorry if it sounded rude, but bob didn't seem to have notice that another had confirmed it, and I was on a mission. Actually yours was the most important post in the thread. And that sentence was part of a very understated thank you.

I thought that was how you meant it.... but, it couldn't resist teasing....

Sonar 6 is using more CPU when a control surface is enabled...

Here are some more findings when a project is loaded and played. This is using the exact same project...

Sonar 5 PE with the control surface enabled:
Task manager reports 74-78% CPU usage for SONARPDE.exe. Sonar reports 59-62 %.

Sonar 6 PE with the control surface enabled:
Task manager reports 96-98% CPU usage for SONARPDE.exe. Sonar reports 54-57 %.

Sonar 6 PE on the same project: NO Control surface
Task manager reports 72-74% CPU usage for SONARPDE.exe. Sonar reports 54-57 %.

CPU usage is NOT the same between Sonar 5 PE and Sonar 6 PE. As reported by windows task manager.

Bob,

I am wondering if you have reproduced this yet. As you can see, it has been reproduced by KEVO and DUOJET.

Thank you
Don

Seth
The only thing I can figure from here is that you system is powerful enough to mask the increase.
I checked the date on my CS dlls on Saturday, they are all 9/11/2006. I also deleted them all and ran a Program ONly install to make sure.
Don

ORIGINAL: xackley
CS makes the Right hand light on the Sonar Task bar MIDI Activity Indicator light up when nothing is happening.

Just a guess, but perhaps the control surface driver is doing some sort of polling, or the control surface itself just contually sends position data. Is the Right Side light input to the PC or output.

In other words, the "culprit" might be the control surface, its driver if it has one, or Sonar's implementation of CS / ACT.

I never noticed the increase in CPU until I saw your thread and checked... I posted what I believed to be useful information that could be used help track this down... There is definitely a substantial increase in CPU usage with a control surface in Sonar 6, as opposed to Sonar 5.

But, the attitude appears to be Micro$oft's standard operating policy... which is... we write the code, and you upgrade your hardware to run it...

It doesn't appear this thread is getting anywhere beyond this....

ACT is no different from how surfaces worked before. The exact same work gets done when the surface is polling for values. The extra work is done once when you click an FX or Synth to change the ACT mapping. Once that is set up (takes less than 1mS), the surface APIs are running exactly the same code they always did. I'm still looking into this, but just wanted to say that ACT or non-ACT parameters are handled exactly the same.

edit: To anyone seeing this thread for the first time, it should be made clear that we're not talking about Audio Engine CPU usage. We're talking about a small increase in the CPU meter reading in the windows task manager which is harmless in regards to Audio performance. I still can't reproduce this here or at home but this is probably indicating some slightly increased Windows messaging going on (in the UI thread, not the Audio Thread(s)). Seeing changes when you drop a menu makes me even more sure of that theory. Us old-timers may remember that up til a few versions ago, because of how we handled the now-time updates, SONAR would always show 100% CPU in the windows task manager (there is still an INI variable you can set if you liked that behavior ). This had no effect on system performance at all. What I'm trying to say is... I'll look into why there is a difference between 5 and 6, but in the meantime, a few percent in the windows task manager is absolutely nothing to worry about.

ORIGINAL: xackley

Bob said in an earlier post that the increased CPU could be because of all the ACT. So I guess an obvious question is Can I shut off all the ACT stuff. It is of no use when running the Mackie Control Surface, as the Mackie CS doesn't support ACT.
Also, why is ACT eating CPU when there is no Learning going on. I could control Sonar from my PCR-50 before ACT using Memory 1 and the Generic Surface with no overhead. I could also control plugins and sonar by mapping controls without the Generic Surface. Once ACT maps (Learns), there is no new functionality that should be using a lot of CPU cycles.
ACT is not important enough that it should be allowed to cripple my machine. I want to just shut it off.

Don

See my "edit:" section in my previous post.

ORIGINAL: xackley

Bob
Have you replicated it on your machines is my biggest question and do you view it as a problem

The CPU meter in a DAW measures the workload on the Audio Engine only. It basically measures how long it took to push a buffer of audio of a given size through the engine. For example, if you have your latency set at 10mS and it takes 5mS to push a buffer through the engine, the cpu will read "50%" If it took 9mS to push the buffer, it would read 90%.

ORIGINAL: John T

Hi Bob,

I'm not having any problems myself, but your description of what the CPU measures has sparked my interest. If you have time, could you explain this in a bit more detail?

I suppose what I'm wondering is... obviously, we can have high Task Manager CPU usage but still have 0% usage in Sonar. Is the Sonar meter measuring a percentage of what remains of CPU capacity? So that 50% will mean different things at different times?

Bob,
Thank you for looking into this...
At least we are reassured that it is being checked into.

Don't know if this is related or not....
Sonar 6.0.1 crashes when changing the output setting in Sonar's control surface dialog.
Select a different midi out other than the one for your control surface. This may or may not crash Sonar on the first time. If you are lucky and Sonar did not crash, select another midi output port again other than the one for your control surface. Crashes every time on my system.

You do not need to close the Control Surface dialog in Sonar when doing the above test...
Windows does generate an error report for this...

I didn't notice this before 6.0.1 update, only because I didn't test it in 6.0.
Previous versions of Sonar do not display this behavior...
I still have 4 PE and 5 PE installed on my system.

I am using the Tranzport...

Bob

Your last post lead me to test putting a project to borderline, and there was NO difference in how sonar reacted with or without a Control Surface configured.

The minor irritation is that now I have to feel guilty having Sonar run all the time. Now I have to think about electricity and heat from the cpu riding at 10-16 instead of 0 to 3, when sonar is just waitng for inspiration.

Thank you for your response

don