[mythtv-users] How MANY backend threads!?!?!? ***** AUTO-REPLY *****

[Raymond Wagner]

On 10/14/2010 10:13, Brian Wood wrote:
> On Thursday, October 14, 2010 08:02:04 am Greg Oliver wrote:
>> On Wed, Oct 13, 2010 at 9:38 PM, Brian Wood<beww at beww.org>  wrote:
>>> On Wednesday, October 13, 2010 07:22:46 pm Greg Oliver wrote:
>>>> Our office was right behind Ti in Dallas.  For the longest they tried
>>>> to crank out compatible x86 CPUs (long before AMD)..  I probably still
>>>> have a few egg trays of those somewhere..  Wish they would have taken
>>>> off..  3 competitors always better than 2..
>>> Well there was Cyrix, and VIA still makes some CPUs.
>> I forgot all about cyrix and never really gave via a chance since
>> their chipsets were just about the worst on motherboards you could
>> find..  Too many memories now..  RMA hell  :)
> I have several thin clients using VIA C7s and they have worked very well, for what they were designed for. I think they are better than Atoms at similar speeds. I think a reasonable frontend could be built with a C7 and VDPAU-capable graphics.

The Atoms are more powerful than the C7s at a per-clock basis, which 
says quite a bit about the C7 that it is bested by an extremely simple, 
in-order core.  The Nanos are in turn more powerful than the Atoms, as 
they should being an OOO core, but they also have a much higher power 
consumption.  The whole lot of them get absolutely demolished by even 
the low end Athlon/Phenom/Core chips, and when you consider a dedicated 
frontend should be off or in standby for most of the time, such a low 
end client is only going to save you a couple dollars per year in 
electricity.

I recall there being something that prevented use of modern nVidia cards 
with VIA hardware.  They have required SSE support for several years 
now, but VIA chips have that capability.  Not sure what I'm thinking of...