| len-dt | Anyone know about usb interrupts? I show 5 interrupts for usb1 to usb5. I have 6 usb ports(plugs)... all of which seem to use usb1. usb1 is ehci (usb 2.0) and usb 2 to usb 5 are uhci (usb1.1 I think) | 01:39 |
|---|---|---|
| len-dt | The MB manual does not show any other usb pins. Odd... anyway, USB seems to be my high priority interrupt. | 01:43 |
| len-dt | eth0 was sharing the same interrupt, but I moved it to be lower than my audio cards and have my ice1712 higher than the ensoniq. | 01:45 |
| len-dt | So far it seems stable. I'm running Audacious through PA->jack at -p 64, downloading software and can run some midi stuff at the same time with no xruns. | 01:49 |
| len-dt | astraljava, todays respin seems to work well live/installed on both my desktop and my netbook and I have marked the i386 as passed in both live and install. | 02:36 |
| astraljava | len-dt: Excellent news, thank you for your efforts! :) | 03:38 |
| len-dt | astraljava, whens the next respin? | 03:41 |
| astraljava | I don't know, hopefully there is none. :) | 03:41 |
| astraljava | I'll sniff the air on -release and let you know if something's happening. | 03:42 |
| len-dt | astraljava, lots activity in #ubuntu-release... | 03:42 |
| astraljava | Heheh, yeah I do fear so... | 03:42 |
| astraljava | Well anything during the last 12 hours doesn't seem to touch us, so thus far I doubt there's going to be a new one on Sunday. But there's still plenty of time left for that. :) | 03:43 |
| len-nb | release day/hour=? | 03:45 |
| astraljava | 26th. | 03:51 |
| astraljava | Why oh why oh why can't people read anymore? | 03:51 |
| astraljava | No offense to whoever that dude is, but I explained all that in the original email. | 03:52 |
| astraljava | Apparently it was a case of TL;DR. | 03:52 |
| astraljava | grr | 03:52 |
| len-dt | astraljava, hey, if he has an amd64, maybe it will get done and then he will know for next time. | 03:55 |
| astraljava | Yeah. :) | 03:56 |
| astraljava | I'd have to start studying how one creates installer packages for Macs. But I think I'll try and get a few more hours of sleep, first. | 03:57 |
| astraljava | Bah, Mail.app doesn't seem to play very nicely together with Gmail. | 04:04 |
| ailo | len-dt: usb should be no problem as long as you don't use it | 06:49 |
| ailo | I haven | 06:50 |
| ailo | 't tried messing with mine, but I have share ir's with usb ports and they aren't affecting performance for me anywyay | 06:50 |
| ailo | wifi and graphic card are commone to cause problems | 06:51 |
| ailo | With the initirq script, shared irq is supposed to be less of a problem, but I'm not so sure | 06:51 |
| ailo | I don't know if we have got any confirmation on this, when it comes to -lowlatency | 06:52 |
| ailo | As I've said before, I don't think you can trim general performance by tuning either the desktop or the kernel | 06:53 |
| ailo | If you get problems, they are most probably hardware related | 06:53 |
| ailo | By tuning the kernel, I mean by switching things off. Of course, the low latency stuff is what matters | 06:54 |
| ailo | len-dt: If you like, try disabling the initirq scriptto see how that affects your performance. It's a dependency to -lowlatency, so what I did was I just moved /etc/init.d/rtirq | 07:24 |
| ailo | len-dt: Also, use "ps -eo rtprio,pri,ni,comm" to see the difference in rtprio for your processes | 07:29 |
| ailo | You'll need to reboot in between, of course | 07:29 |
| len-dt | ailo, sorry I had gone to bed (not beauty sleep, just non-ugly sleep) | 14:32 |
| len-dt | I found that my pci slots starting right after the video slot go from low to high. | 14:33 |
| len-dt | The last one is int 23, but is shared with all the usb2.0 ports. I had the nic in there for that reason, but I found out that hardware wise it has priority... so net traffic affected my sound. | 14:36 |
| len-dt | I have moved the nic to beside the video card... it gets int17 now. | 14:38 |
| len-dt | The end slot where it was is empty to avoid int sharing with a sound card and the one next to that (int 22?) still had the knockout so I left it empty too for now. Next is the ice1712 with int 21. It seems to be rock solid now... -p32 is possible with the odd xrun, but -p64 only gets xruns when PA's jacksink closes a port... not jack's fault and not a good low latency use anyway. | 14:42 |
| len-dt | next the ice1712 is the ens1370 with int20. It gets used for midi io only and the audio part just causes trouble. | 14:50 |
| * len-dt wishes he could turn that part off and just get midi. | 14:50 | |
| len-dt | I have to make sure I turn it off in PA after a new install or it can cause xruns... maybe not so much now that the int is lower priority. | 14:53 |
| len-dt | I have found that turning off pulse->jack bridging brings cpu usage down a lot in -p128 and lower. Try: | 14:57 |
| len-dt | pactl unload-module `pactl list short |grep jackdbus |sed s/module-jackdbus-detect//` | 14:57 |
| len-dt | To test this. and pulseaudio --kill to restart. | 14:58 |
| len-dt | At -p1024 the effect doesn't show, but lower, PA's CPU goes up about twice as fast as jacks. | 15:02 |
| len-dt | I have turned off the par and ser ports in bios and found out I have ir and firewire (also turned of) as well. | 15:11 |
| len-dt | ailo, ps -eo rtprio,pri,ni,comm, gives this: | 15:16 |
| len-dt | 85 125 - irq/20-snd_ens1 | 15:16 |
| len-dt | - 19 0 dnsmasq | 15:16 |
| len-dt | 85 125 - irq/21-snd_ice1 | 15:16 |
| len-dt | should I be running the ens at 80? | 15:17 |
| len-dt | also I have: | 15:18 |
| len-dt | 80 120 - irq/23-ehci_hcd | 15:18 |
| len-dt | 80 120 - irq/16-uhci_hcd | 15:18 |
| len-dt | 78 118 - irq/19-uhci_hcd | 15:18 |
| len-dt | 79 119 - irq/18-uhci_hcd | 15:18 |
| len-dt | 79 119 - irq/16-uhci_hcd | 15:18 |
| len-dt | But with no usb devices that should not be a problem. | 15:18 |
| len-dt | I will try testing with that stuff turned off to see what difference it makes | 15:50 |
| len-dt | ScottL meeting today? | 16:28 |
| ScottL | len-dt, we were supposed to | 16:29 |
| ScottL | we still can | 16:29 |
| ScottL | i'm doing a few errands around the house but i'll poke in quite often and see if we reach a quorum for a meeting | 16:29 |
| len-nb | ScottL, I have to go in the next few minutes. | 16:36 |
| astraljava | So you guys come up with a schedule. I'll probably go to sleep within the next couple of hours, so if it happens during that time, I can participate. | 16:42 |
| astraljava | But we seriously have to commit to meetings better the next cycle, | 16:43 |
| astraljava | It's been rather ridiculous for the current one. | 16:43 |
| len-nb | astraljava, ja, no kidding | 16:43 |
| len-nb | I get the idea people are not yet ready to talk about next cycle, being worn out from this one. | 16:44 |
| astraljava | len-nb: Yeah, and it's why I'm not trying to squeeze anyone for best times for it, yet. | 16:45 |
| astraljava | Let us all cool down for a moment. :) | 16:45 |
| len-nb | astraljava, I think so, also maybe the time change may confuse people. | 16:46 |
| len-nb | I have gone from -9 to -8 | 16:47 |
| astraljava | Yeah. But then it also seems this current schedule isn't the most optimal. | 16:47 |
| len-nb | Anyway, I have to go. | 16:49 |
| * astraljava laughs at the mail on -users | 18:29 | |
| astraljava | Is someone actually working on the Preparation page? | 18:29 |
| astraljava | We should remove the "under construction" statement if not. | 18:30 |
| ScottL | astraljava, i believe ttoine was doing it | 18:35 |
| astraljava | Ok. | 18:37 |
| === jussi is now known as android | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!