[haiku-development] Re: Res: Scheduler algorithm improvements

  • From: Cian Duffy <myob87@xxxxxxxxx>
  • To: haiku-development@xxxxxxxxxxxxx
  • Date: Sat, 7 Nov 2009 16:49:49 +0000

2009/11/6 Urias McCullough <umccullough@xxxxxxxxx>:
> On Fri, Nov 6, 2009 at 10:43 AM, Cian Duffy <myob87@xxxxxxxxx> wrote:
>>> I have read some people having problems with certain hardware - I have
>>> always thought these are interrupt-sharing related issues, or a bug in
>>> the driver perhaps. Not scheduler-related at all.
>> All my working machines have HDA, unfortunately for checking this out;
>> but I still have serious "spitching" problems with audio under load.
>> One needs OSS, one doesn't, both have this problem though.
> Strange...
> I know on my AA1 for example with HDA - running the Big Buck Bunny
> video (480p) while loading up the CPU doesn't make audio glitch (at
> least not until I add enough load to make the video glitch as well).
> Perhaps it would maybe be worth aggregating some useful information (i
> don't know which info, and please not in this thread) among people who
> have issues and those who don't to see if there's a pattern that can
> be used to track down the cause.

To follow up on this, on the machine with a HDA that needs OSS I've
put a very aged Creative EMU sound card in its single available PCI
slot (mini-ATX+massive graphics card) and the audio glitching is down
to maybe one drop out every few minutes under load rather than ten a
minute. This is a native driver, not OSS.

Unfortunately the other machine that uses a HDA with the native driver
is an Acer Aspire like your own (albeit the very bog standard model -
you might have something fancier?) and can't take a replacement card,
but it does have the glitching issues also.


Other related posts: