• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: MOTU Fastlane MIDI driver wastes 10-25% CPU
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MOTU Fastlane MIDI driver wastes 10-25% CPU


  • Subject: Re: MOTU Fastlane MIDI driver wastes 10-25% CPU
  • From: Herbie Robinson <email@hidden>
  • Date: Sun, 10 Nov 2002 01:35:03 -0500

On Friday, November 8, 2002, at 03:04 PM, Doug Wyatt wrote:

On Thursday, Nov 7, 2002, at 23:23 US/Pacific, Kurt Revis wrote:
What I DO get are periodic spikes where the CPU usage goes up to 100% for no apparent reason. This goes on for around fifteen seconds, and happens every few minutes (it's happened 3 or 4 times while I've been writing this). The MIDIServer's CPU usage bounces around between 50% and 100%; most of this is shown as system time, not user.

Please open your console.log and /var/log/system.log and tell me whether these periods of high usage correlate with log messages from /usr/libexec/fix_prebinding.

Nope, I would have noticed that (although I did check again). Top shows the CPU being used by the MIDIServer, not fix_prebinding.

When I run sample on the MIDIServer, here's the only interesting thread:

<SNIP>
36 IsocUSBRead::DoRead(unsigned long long)
36 IOUSBInterfaceClass::ReadIsochPipeAsync(unsigned char, void*, unsigned long long, unsigned long, IOUSBIsocFrame*, void (*)(void*, int, void*), void*)
36 io_async_method_structureI_structureO
36 mach_msg
36 mach_msg_trap
36 mach_msg_trap [STACK TOP]

However, this looks exactly the same whether or not the CPU is spiking, so the driver doesn't appear to be directly at fault.
Also, running Shikari during a spike shows the MIDIServer at about 5% usage (pretty negligible) and the kernel taking up 70% of the time, sitting in machine_idle_ret. But again, running it when there's no spike gives essentially the same results.

So I suspect there's nothing really bad going on, just some sort of weird anomaly in the kernel that causes it to erroneously report high CPU usage. (Both top and the CPU Monitor app show the spike, BTW.) But I could very easily be wrong!

Oops, ReadIsochPipeAsync shouldn't be blocking in the kernel. Looks like this isn't MOTU's bug...
--
-*****************************************
** http://www.curbside-recording.com/ **
******************************************
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.
References: 
 >Re: MOTU Fastlane MIDI driver wastes 10-25% CPU (From: Kurt Revis <email@hidden>)

  • Prev by Date: Re: Use of the source refcon in MIDIProc callback
  • Next by Date: Re: MOTU Fastlane MIDI driver wastes 10-25% CPU
  • Previous by thread: Re: MOTU Fastlane MIDI driver wastes 10-25% CPU
  • Next by thread: AudioUnits...
  • Index(es):
    • Date
    • Thread