[mythtv-users] Superserver

Mark Hutchinson markhsa at gmail.com
Tue Aug 24 20:09:12 UTC 2010


Thanks.  That makes good sense.

On Tue, Aug 24, 2010 at 2:02 PM, Michael T. Dean <mtdean at thirdcontact.com>wrote:

>  On 08/24/2010 02:51 PM, Mark wrote:
>
>> Mark Hutchinson wrote:
>>
>>> Good point yes.
>>> What might some good options be for small frontend machines be that can
>>> be turned off and on easily? I plan to have the frontends all in the
>>> basement as I have 3 CAT 6 cables run to each TV.  2 cat6 for HDMI and the
>>> 3rd for a remote or kb/mouse.
>>> Does suspend work for this?  How would they be woken up?
>>>
>>> Thanks for the thoughts.
>>>
>> have you considered mini-ITX atom boards for local frontends?  They are
>> very small and make no noise.
>>
>> Remote pc's and long wires are a pain, in case you have'nt played with
>> that yet...
>>
>
> OK, I'm not going to recommend any specific system.  Normally, I wouldn't
> even reply, but I'd just like to inject a bit of non-marketing reality into
> the thread.
>
> Note, also, in the interest of full disclosure, I am a /confirmed/
> Atom-hater.  I have a huge and deep bias against Atom.  (For some reason, I
> feel a computer should be able to compute.)
>
> That said, low-power doesn't have to mean a toy.  See what proper design of
> a real computer system can do (whether you do it or Apple does):
>
> http://www.apple.com/macmini/specs.html
> http://support.apple.com/kb/HT3468
>
> (10W idle and 85W max power consumption) and if you figure you'd be using
> VDPAU for decoding /when VDPAU works with the stream you're decoding/, you'd
> be running at close to that idle 10W even when decoding, for example, h.264
> video.  The big difference, however, is that you actually have computing
> power in reserve when you need it (for Flash or video that doesn't meet the
> codec, profile, bitrate, and deinterlacing limitations of VDPAU or whatever
> other proprietary problems we can't solve).
>
> Now, even if an ION system ran at 0W, and we assume the Mac Mini is running
> at close to idle when using VDPAU decode, that's a typical savings of only
> about 10W.  And, if you're shutting down your frontend when not in use, the
> difference between an Atom-based toy and a real computer in power usage is
> not a lot.  The difference in performance/limitations, however...
>
> To put some numbers to it, let's use the preposterous assumption that an
> ION system runs at 0W under full load and the Mac Mini (or a system designed
> like it) runs at the full 85W all the time (under load or idle).  Further,
> let's say you shut down your frontends when not in use (as recommended
> above) and you use them for about 2hrs/day.  That means that the Mini is
> using 85W * 2hr (= 170Wh) per day.  That's 5100Wh/mo = 5.1kWh/mo.  Assume a
> national average retail price of $0.12/kWh (
> http://www.eia.doe.gov/electricity/epm/table5_6_a.html ), that's $0.612
> per month to run the Mac Mini/$0.612 saved by running an Atom system.  Now,
> when you factor in the actual power usage (which will nearly always be less
> than--and likely /significantly/ less than--the 85W max), the Atom-based
> systems may not make much sense.  Basically, it comes down to some price
> between $0.0012/hr (at 10W consumption) to $0.0102/hr (at 85W consumption)
> to run the Mini--meaning the maximum savings you could get from running an
> Atom-based system is a penny an hour.
>
> Basically, the saving from dropping your computer power usage from 170W to
> 85W is /much/ more important than the savings from dropping your computer
> power usage from 85W to 42.5W (even though in both cases, the power usage
> was halved).  When you get down to dropping from, say, 20W to 10W, there's
> not much savings involved.  (It's the same as fuel efficiency--trading your
> 10mpg vehicle for a 20mpg vehicle is much more beneficial than trading your
> 20mpg vehicle for a 40mpg vehicle.)
>
> See, also, these (old) articles:
> http://www.tomshardware.com/reviews/25w-performance-pc,2551.html
> http://www.tomshardware.com/reviews/intel-e7200-g31,2039.html
> http://www.tomshardware.com/reviews/intel-atom-efficiency,2069.html
>
> I'm sure that there are newer articles using newer chips/architectures that
> do even better.  IMHO, Atom is a marketing ploy that allows Intel to charge
> significantly more while providing no advantages and only disadvantages.
>
> OK, sorry.  I'm stepping down off my soap box, now.  I won't reply to this
> thread as I'm not trying to start a big war with Atom fans.  So, if some
> Atom fanatic needs to get the last word, feel free to reply, and I won't
> dispute anything you say.  If this post doesn't stand on its own, maybe it
> will at least get some people to think a little deeper about the issues--dig
> down beneath the marketing.
>
> Good luck with your new setup, and I hope you find a way to reduce power
> consumption without losing capability.  And I'll reiterate that the /best/
> way to do so is to shut down systems when not in use.
>
> Mike
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mythtv.org/pipermail/mythtv-users/attachments/20100824/e5dc42fe/attachment-0001.htm>


More information about the mythtv-users mailing list