On Mi, 2.03.2005, 16:41, [EMAIL PROTECTED] said:
>
> Hi,
>
Hi again!

> I was having similar problems on (almost) the same hardware.
> I'm running Debian with a self-compiled 2.6.10 kernel, 2 PVR-350 cards,
> software RAID5,
> Asus P4P800 Deluxe, Intel P4 3,2GHz Prescott but with HT DISabled.
>
Doh! Unbelievable! GRMPF!
It works... flawless... for more than 5 hours with very heavy channel
switching, ffw, rewind (timeshifting), watching recordings, etc.
I even stressed the CPU with doing some transcoding and even a
[EMAIL PROTECTED]
LiveTV runs smooth, no glitches, no artifacts...
ARGH!!!

> With HT enabled, when both PVR's were busy (recording, watching live TV,
> playing a
> recording, whatever), Myth would freeze within 60secs. sometimes even
> needed to reboot to
> get it running again...
>

I was able to record as hell. Several hours on both cards simultaneously.
But as soon, as I started to use the hardware supported output in Mythtv
it startet to freeze within a few minutes.


> I was receiving exactly the the same error messages (ENC: Sched DMA) in
> the logs...
I dont think it is a error-msg.
I think it just tells you, that a DMA-Transfer was scheduled.

>
> After trying for 2 weeks weeks to get the box to run stable (the thought
> of window,
> computer and throwing out also crossed my mind ;)), I decided to try and
> disable HT.

Because I bought the more expensive Intel Stuff, for having SMP and (as i
thought) better load balancong on the machine, I did not even think about
this. I was sitting there, several time just thinking "This has to work
WITH Hyperthreading!!!"

> Without HT, I now have an uptime of 9 days, using only Myth to watch TV
> (while doing
> multiple simultaneous recordings etc). No lockups at all anymore!
>
While I updated my Debian Sarge DVDs yesterday evening I was sitting
there, watching for 6 hours and using timeshifting and everything... This
was not good to my heart... I'm getting too old for this...


> Looks like someting in the current DMA logic is not happy with either SMP
> or with the way
> HT is implemented as SMP (perhaps the P4P800 has something to do with it?)

As there seem to be many people out there (i even know 3 of them), having
PVR350's running on P4 with HT, but not on this Mainboard, I would guess
the DMA-Logic in BIOS (has BIOS really to do with this?!?) or the way the
DMA-Controller is used/linked (i don't think the way of linking the
DMA-Controller is relevant, as it is included in the Chipset, so i think
it's some kind of Software problem) is, to speak clearly really fucked up
on this board.
Correct me, if someone knows better... ;)

> Mind you: I'm using exactly the same (SMP)kernel, I only disabled HT in
> the BIOS.

I'm going to build a 2.6.10 non-SMP kernel in the evening today and setup
everything, as i want and then inform you, what happened.


>
> I'm also using the latest vsync patch (still have a minor issue that I
> hoped this patch
> would fix, but didn't), haven't tried what the effect of this patch is on
> the HT/SMP
> instability...

What does this patch do? Should i use a 0.3.x Version of ivtv, or should i
stay with the 0.2.0-rc3g? D'oh! I'm on the ivtv-Devel-List...
Idiot-Question I think...
>
> Anyone else have similar problems or other experiences on SMP and/or HT
> P4's?
As A.P. Kennedy reports, that HT works for him, and recalling the fact I
know 3 guys using P4 with HT without any problem, I think it is either
hardware f'cked up and/or a BIOS/Software problem related to the ASUS
P4P800 with HT enabled.


>
> Regards,
> Stanley.
>
Thanks, for your quick reply,

Jan Ophey

--
Reality is, where the Pizzaguy comes from...



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
ivtv-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ivtv-devel

Reply via email to