Re: [LAU] hardware - Intel CPUs

From: david <gnome@email-addr-hidden>
Date: Fri Apr 18 2014 - 10:08:03 EEST

On 04/17/2014 09:53 AM, Len Ovens wrote:

> On Wed, 16 Apr 2014, david wrote:
>
>> On 04/16/2014 03:17 AM, Len Ovens wrote:
>>> I was talking about through air. I can only play as fast as I can hear
>>> and at thirty feet or so from the sound source, my instrument sounds
>>> delayed from the rest of the band because I hear their sound that much
>>> later and play my part that much later. I don't normally play at that
>>> distance, but ten feet is pretty standard.
>>
>> You're a few orders of magnitude better than me! I don't think I'd
>> notice that much difference.
>
> I used to play and try to adjust sound while practicing, so I had 50
> feet or so of cable. By the time I got to the sound board it was getting
> hard to play.

I think I would, too. When I was playing bass on stage ages ago, I don't
think I got more than 20 feet from my amp. I wasn't a singer, so no
need. The one point where we needed cable was when the lead guitarist
and I would finish one song by somersaulting across the stage while playing.

My church band's former music director used to wirelessly hook his
guitar into the system, then walk around while we rehearsed so he could
hear how things sounded in the auditorium. That was in the days when we
could get the place whenever we wanted to set everything up for
rehearsal. Our sound tech has gone around with a soundmeter in our new
location and fiddled with levels. And if one of us has to miss practice
(our rule is you don't get to play that Sunday), we try to sit various
places and afterwards talk about how things sounded.

>>> However, it looks like I can still get lots of MB with PCI slots in
>>> them. I will probaly do that. Hopefully with three PCI slots I can get
>>> one that is irq clean.
>>
>> My desktop has 2 PCIe slots; the other 4 are PCI. It also has 4 USB
>> ports + 2 more USB connector points on the mobo, 4 SATA connectors, an
>> EIDE connector, a floppy connector, plus the built-in audio and video
>> and Ethernet. I haven't checked to see what's sharing interrupts and
>> what isn't. I wouldn't be surprised if there aren't a bunch of shared
>> interrupts!
>
> cat /proc/interrupts will tell you that quick enough. For some reason a
> lot of MB share irq16 with 3 or 4 things... worse one of those tings is
> often one of the PCI(e) slots. No need really as most modern MB have
> access to 48 plus irqs. My old board has 24 but 2-7 are unused (for
> hysterical reasons?) and 10 and 13 seem to be skipped too. I think 20
> goes to my internal audio which I have turned off (AC97) so it doesn't
> show.

I think mine has a clump on IRQ16. I thought leaving behind the old IBM
AT bus architecture was supposed to free us from IRQ problems?

> I have noticed that on any of the MB I have checked or seen irq maps for
> that the internal audio always has it's own irq and quite often it is
> the highest one which in most systems these days has the most priority.
> I found this out because I had the card I was using for midi in the
> higher of the two slots and my audio below. I had trouble with xruns on
> the audio, but when I put the audio on the higher irq I had no more
> trouble.

I've discovered on my desktop that if I disable the onboard audio (which
I don't use at all), ALSA doesn't load or start. But that motherboard on
that is starting to have problems - sometimes it will boot up with a
blank screen for a moment or so, then the BIOS display appears with a
red warning that the "last overclocking setting failed". Which is odd
because I have never and am not now overclocking it. I guess if it
finally goes out I'll replace it with an Intel board. I like the Haswell
architecture, it gives my laptop great performance while maintaining
battery life (about 3.5hrs with my 2.4GHz i7). The previous model used
slower Intel chips to squeeze out that battery life. Now if only video
support for the HD4600 chip gets better - playing MP4s or other video
formats results in frozen image except for a band or two that change.
Ubuntu 13.10 can play video on it (if you install a bunch of libraries
from outside the Ubuntu reality), so maybe it's just taking some time
for Ubuntu's X tweaks to get back into the general X codebase.

> I know that in theory that shouldn't happen because there are two part
> to the irq drivers, a stub to answer the IRQ and save enough info to
> work and the other part that the os prioritizes and does all the work.
> So the os should be able to prioritize by the module name. I just know
> what I have found works best.
>
> I have heard the words "in a modern system" too many times. I think any
> system can do better audio if it is tuned/tweaked.

Sure! I even had decent audio working on my old effectsbox laptop, which
was an old Toshiba with a 2.8GHz Celeron and only 768MB of memory. It
was old before it landed in my collection! (Audio quality was fine.
Processor, on the other hand, was not really up to the job!)

Toshiba laptops used to be really easy to use with Linux; Toshiba used
"bog standard hardware" (as a friend of mine called it) and everything
had full, mature Linux support. Don't know about current Toshibas.
They've probably been pulled further into the Windows orbit since then.

-- 
David W. Jones
gnome@email-addr-hidden
authenticity, honesty, community
http://dancingtreefrog.com
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user
Received on Fri Apr 18 12:15:02 2014

This archive was generated by hypermail 2.1.8 : Fri Apr 18 2014 - 12:15:02 EEST