BrainModular

Troubles with High CPU and Latency

2013-11-05T20:24:11+02:00

You might be confusing the irq number and the priority, i don't believe those are directly related as in irq 1 has higher priority than irq 2.

You can do some testing and find out which usb port on your computer is the "1E26" on iRQ23 and use this one for the soundcard, it might be better than a usb port on irq 16 shared with the graphics.
You disable one usb device in the Device Manager and then you open msinfo32 and you can see the one that has disapeared.
in the end you can match the physical usb port to the device in the manager to the device in the RQ list and then use the priority trick to give high priority to the soundcard usb.

if you end up using the usb port on irq 23, Then you can set up a high priority on iRQ23 and low priority on iRQ16 for example.
That's how i did on my laptop.

Also some laptops with intel sandy bridge, ivy bridge cpus have an internal graphic thing, so that's in fact two graphic devices (intel HD graphics 3000 + Nvidia GeForce something for example on two different irqs !!!).

Statistics: Posted by FarfadetFarfelu — 05 Nov 2013, 19:24

Troubles with High CPU and Latency

2013-11-05T11:14:33+02:00

I was looking at the irq's on my machine and was a bit confused. Maybe someone knows more than I.
Im trying to find the irq related to my usb devices. My graphics card has the same irq as my usb

IRQ 16Intel(R) 7 Series/C216 Chipset Family USB Enhanced Host Controller - 1E2DOK
IRQ 16NVIDIA GeForce GTX 675MXOK

I also have2 other related to usb
IRQ 23Intel(R) 7 Series/C216 Chipset Family USB Enhanced Host Controller - 1E26OK

IRQ 4294967293Intel(R) USB 3.0 eXtensible Host Controller - 0100 (Microsoft)OK
/?
how would I prioritize the first usb irq if it is the same as the graphics card?

IRQ 0System timerOK
IRQ 1PC/AT Enhanced PS/2 Keyboard (101/102-Key)OK
IRQ 8System CMOS/real time clockOK
IRQ 12Synaptics PS/2 Port TouchPadOK
IRQ 13Numeric data processorOK

also it looks as if my machine prioritizes the touch pad and keyboard over the numeric data processor. strange!!

any recommendations?

Statistics: Posted by seamus — 05 Nov 2013, 10:14

Troubles with High CPU and Latency

2013-11-05T02:09:06+02:00

I notice you set a higher priority for foreground applications, but I've always been taught that audio streaming was a background process, and you should be giving background processes higher priority (which is a simple switch in the system preference tab)..... that's been in every 'audio optimization' page I've ever seen.

Is this a misconception?

Statistics: Posted by woodslanding — 05 Nov 2013, 01:09

Troubles with High CPU and Latency

2013-11-05T00:44:19+02:00

here : (and it works with win .

http://www.sensomusic.com/wiki2/doku.ph ... q_priority

basically i had two big master audio vu meters on my patch that were really making everything click early.
So yeah i really wonder if this could be further optimized, it seems strange that they cause so much trouble !

Statistics: Posted by FarfadetFarfelu — 04 Nov 2013, 23:44

Troubles with High CPU and Latency

2013-11-04T23:35:14+02:00

I'm trying to find that irq priority tip, and not turning it up..... can you send a link?

I have a lot of VU meters in my wkp, that are attached via audio busses. I'm wondering if they could be data busses, and if so, would that decrease the processor load? Seems like overkill for a vu meter to be an audio signal, as they can only refresh at the video refresh rate, correct? I could take them out altogether, but they are helpful.

Or is there a more proper way to 'down-sample' an audio signal???

Statistics: Posted by woodslanding — 04 Nov 2013, 22:35

Troubles with High CPU and Latency

2013-11-04T21:33:55+02:00

Same here, but optimizing my laptop really helped.
The crackles appeared to be graphic related as closing the interface builder stopped the crackes.

- First i checked the iRQs and realized the usb port i was using for the soundcard was the same irq as the graphic card !
moved the soundcard to another usb port, huge improvement.

- Then i took out non essential vu meters and other graphic stuff from my patch.

- Then i used the irq priority Tip from the wikipedia to setup the graphic cards to a lower priority... i was amazed it further improved the situation !

Statistics: Posted by FarfadetFarfelu — 04 Nov 2013, 20:33

Troubles with High CPU and Latency

2013-10-18T02:24:18+02:00

Okay well I am trying a simple interface right now, using a sequence line from 0 to 127, rounded and a has changed to trigger a Create MIDI module. The module is set to Controller, and the MIDI is being bussed to a plugin (Z3TA 2).

It seems I am having high CPU outputs even with a small resolution of steps from the line sequencer, as well as I have to pump the driver buffers to 1024 to reduce/remove cracks and pops.

Any suggestions regarding implementation? I cannot believe that I am having CPU problems...and I have never had to push more than 64/128 buffers using my RME HDSP.

For such a simple task I cannot believe I am having so many issues. I can use the mouse or touch and change the faders smoothly with no pops and cracks...only seeing issues where I am feeding controller data .

i had a similar problem of cpu while using a panel , but i have solved the problem closing the non-used (minimized) windows..
maybe loading the default settings in your asio driver could increase performance, i dunno if this could help ,..

PD: also one thing i know , is Zeta+2 is a bit heavy in cpu usage , i know it because i also use it , and i have a x4 cpu Pc..

good luck

Statistics: Posted by Taysete_DJ — 18 Oct 2013, 02:24

Troubles with High CPU and Latency

2013-10-16T03:00:31+02:00

Okay well I am trying a simple interface right now, using a sequence line from 0 to 127, rounded and a has changed to trigger a Create MIDI module. The module is set to Controller, and the MIDI is being bussed to a plugin (Z3TA 2).

It seems I am having high CPU outputs even with a small resolution of steps from the line sequencer, as well as I have to pump the driver buffers to 1024 to reduce/remove cracks and pops.

Any suggestions regarding implementation? I cannot believe that I am having CPU problems...and I have never had to push more than 64/128 buffers using my RME HDSP.

For such a simple task I cannot believe I am having so many issues. I can use the mouse or touch and change the faders smoothly with no pops and cracks...only seeing issues where I am feeding controller data .

Statistics: Posted by sephult — 16 Oct 2013, 03:00

Troubles with High CPU and Latency

2013-11-05T20:24:11+02:00

Statistics: Posted by FarfadetFarfelu — 05 Nov 2013, 19:24

Troubles with High CPU and Latency

2013-11-05T11:14:33+02:00

Statistics: Posted by seamus — 05 Nov 2013, 10:14

Troubles with High CPU and Latency

2013-11-05T02:09:06+02:00

Statistics: Posted by woodslanding — 05 Nov 2013, 01:09

Troubles with High CPU and Latency

2013-11-05T00:44:19+02:00

Statistics: Posted by FarfadetFarfelu — 04 Nov 2013, 23:44

Troubles with High CPU and Latency

2013-11-04T23:35:14+02:00

Statistics: Posted by woodslanding — 04 Nov 2013, 22:35

Troubles with High CPU and Latency

2013-11-04T21:33:55+02:00

Statistics: Posted by FarfadetFarfelu — 04 Nov 2013, 20:33

Troubles with High CPU and Latency

2013-10-18T02:24:18+02:00

Okay well I am trying a simple interface right now, using a sequence line from 0 to 127, rounded and a has changed to trigger a Create MIDI module. The module is set to Controller, and the MIDI is being bussed to a plugin (Z3TA 2).

It seems I am having high CPU outputs even with a small resolution of steps from the line sequencer, as well as I have to pump the driver buffers to 1024 to reduce/remove cracks and pops.

Any suggestions regarding implementation? I cannot believe that I am having CPU problems...and I have never had to push more than 64/128 buffers using my RME HDSP.

For such a simple task I cannot believe I am having so many issues. I can use the mouse or touch and change the faders smoothly with no pops and cracks...only seeing issues where I am feeding controller data .

Statistics: Posted by Taysete_DJ — 18 Oct 2013, 02:24

Troubles with High CPU and Latency

2013-10-16T03:00:31+02:00

Statistics: Posted by sephult — 16 Oct 2013, 03:00