Re: [LAU] kernel changes

On Sun, February 10, 2013 9:37 am, Ralf Mardorf wrote:
> On Sun, 2013-02-10 at 09:25 -0800, Len Ovens wrote:
>> What I am says though is that I'm not sure these problems are kernel
>> problems. They may be kernel module problems or something else. I have
>> used all of these kernel versions and been able to run low latency on
>> all
>> of them no problems (ICE1712 based card). Other people on this list have
>> had no problems too. Yet there is not just one person with this same
>> problem. We (on this list) might do well to explore the differences
>> between our hardware/software to see if we can find the problem.
>
> Hi Len :)
>
> regarding to latency and xruns _on my machine_ two cheap ICE1712 PCI
> cards are working better on Linux, than my RME PCI express card. Tested
> with kernels from 2.6.29 full preempt on Debian to

I am wondering if PCIe or it's drivers may be a problem for low latency.
Having read a little about the standard, it would seem that the earliest
standards (v1) are better for low latency than later (v4) standards. The
improvements have been in throughput at the expense of latency. I wonder
if it is possible to have a PCIe slot set to a particular version of the
PCIe standard. Or to do so at the card end. As a distro assembly
volunteer, I can only test things with the HW I have, which is older (no
PCIe slots at all). What I have read about the RME PCIe products is that
they are native PCIe rather than bridged PCI to PCIe.

I wish I had the funding to test full time and buy a better representation
of the HW out there, but I don't.

-- 
Len Ovens
www.OvenWerks.net
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user