Talk About Network

Google


Register and Login
Nick
Password
Register create new account Sign up is FREE and you can post replies, new topics, bookmark posts and more!
Recover lost password


Electronic Equipment > PC Hardware Chips > Re: A case for ...
Latest [ Topics | Posts ] Archive Post A New Topic Post a Reply
<< Topic < Post Post 3 of 24 Topic 2688 of 2798
Post > Topic >>

Re: A case for Dual System Cabinets

by kony <spam@[EMAIL PROTECTED] > Dec 26, 2007 at 03:58 AM

On Wed, 26 Dec 2007 00:14:42 -0800 (PST), Ht****irs
<ht****irs@[EMAIL PROTECTED]
> wrote:

>Lost angel:
>
>The power figures have been collected from various online sources. And
>15" monitor was a CRT which is still more popular in my country.

Even a low end 25W video card is easily double the
performance of integrated video.  Yes a high powered gaming
video card can consume 75W, but remember that even those
that consume a lot, consume significantly less when not
gaming.


>
>I agree that power utilisation changes with the load. But, the point
>is that even when dealing with less intensive work a lot of power is
>lost, because the components in modern systems are inherently power
>hungry. Which is why there is the need for a power miser system within
>the same cabinet to be used when dealing with loads that do not
>require the power of the "main" system.
>


I disagree.  remember that the Via motherboard chipset is
not especially optimized for efficiency, while the processor
is but because it has such low performance it will be
operating at a substantially higher % of max load
continuously, while other reasonably efficient processors
would operate at very low load and thus mostly idling with
much lower than average or peak power consumed.



>Please inform what the dual SLI system working on when the 250W
>consumption was measured.
>
>Kony:
>Thanks for the CF to IDE converter idea. I was not aware of that
>option, and that converter is not easily available here. Getting the
>system to boot  in USB is admittedly quite complicated when dealing
>with XP. In Linux, it is more easier.

Even then in Linux, it is much slower.  A converter is not a
hard thing to acquire, many of the Chinese/et al
distributers will ****p globally for low total cost.


>I have not yet tried the VIA system I mentioned. Though I did read
>that they are quite weak in comparison with even the older generation
>CPUs of Intel and AMD. VIA Mini ITX products are not available easily
>here, and I am awaiting a response from the main distributor. But,
>reading the spec sheet of the EPIA SN series (http://www.via.com.tw/en/
>products/mainboards/motherboards.jsp?motherboard_id=550), I did
>believe that it would suffice for the work at hand. 


Sometimes yes it would suffice, but then you may come upon a
situation where even when what seems a modest task, is
terribly slow because of low performance ceiling.

There is more performance to be had and a more reasonable
modern compromise by underclocking contem****ary Athlon64 or
Core2 architectures.  Remember we can't look at peak CPU
power if we are talking about lower processing demand
workloads, rather we'd have to assume the higher performing
processors are far more often idling with significantly
lower power usage.

A Pentium 3 based system with optimized component choices,
including a SSD or CF-IDE based drive, can operate off a 50W
PSU.  A modern system with similar concessions including
integrated video can operate off a 100W PSU - if only there
is enough current on the 12V rail.  

Someone doesn't need to make these compromises though,
consider for example a Geforce 7600GT video card.  In
regular 2D use it may consume no more than a dozen watts,
and under 35W in gaming.  If one wishes to further reduce
it's power consumption they have both the easy way and
advanced way to make that happen.  The easy way is use a
software (or bios editing tool to permanently set) for lower
core and memory speeds, which is a linear decrease in power
consumption.  The more advanced way would be to reverse
engineer the power regulation subcircuits to force a lower
voltage, and some testing to determine what the best
voltage:frequency tradeoff is.  Taking above example a
Geforce 7600GT should still be able to offer over twice,
possibly over 3 times the performance of integrated video
while consuming under 15W when runnning in an optimized
voltage:frequency configuration.

The problem with Via alternatives is they force a very low
performance level.  IF that performance level were ok then
nobody would have bought their higher performing system,
they would have just kept using their old system instead of
replacing it with parts that "might" have almost the power
consumption figures you mentioned.


>And the mini ITX
>motherboard makes it possible to  include it in the same cabinet,
>probably screwing it onto a hinged side panel (not slideable). Other
>details need to be worked out.
>
>By the way, which model of Via did you have problems with?
>
>

CLE266 based PCChips board with Via C3 processor.  I forget
the model number.  While the performance was so low I would
definitely not use it for regular tasks, I had hoped to use
it for a fileserver but after all kinds of driver hacks,
bios changes, and even direct chipset register changes via
utilities, it still couldn't provide normal PCI throughput
without corrupting data.

I think in a few years Via chipsets plus CPU boards will be
a good choice.  After they learn a bit from mistakes and
most things sit on the PCI Express bus or are southbridge
integrated.    Even then, modern alternatives from other
sources still have a good tradeoff of only slightly higher
power consumption with much more performance potential if
only one does not run that alternative at the max frequency
and voltage possible.

Many systems of yesteryear can be overclocked and
undervolted to result in surprising decrease in power used,
and even better, it can be done with no addt'l cost to buy
new equipment, no new cost for another OS license (if it
runs windows), and no manufacturing pollution or landfill
waste from throwing away old parts to replace with low
performing newer ones.

There is likely some system within 100 mi. of anyone who
wants one, which was either abandoned or sold cheap because
it has low performance like a newer Via CPU/chipset based
system would.
 




 24 Posts in Topic:
Re: A case for Dual System Cabinets
"philo" <phi  2007-12-25 09:59:00 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2007-12-25 04:42:40 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-26 03:58:28 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2007-12-25 15:24:51 
Re: A case for Dual System Cabinets
Robert Redelmeier <red  2007-12-26 23:51:19 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-26 23:59:33 
Re: A case for Dual System Cabinets
Robert Redelmeier <red  2007-12-27 05:32:38 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-27 02:06:52 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-27 14:16:03 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2007-12-27 03:51:41 
Re: A case for Dual System Cabinets
Sebastian Kaliszewski <  2007-12-28 12:56:58 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2007-12-28 13:34:35 
Re: A case for Dual System Cabinets
Sebastian Kaliszewski <  2008-01-02 15:17:50 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2008-01-03 18:22:27 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2008-01-03 17:25:33 
Re: A case for Dual System Cabinets
a?n?g?e?l@[EMAIL PROTECTE  2008-01-04 16:33:19 
Re: A case for Dual System Cabinets
Sebastian Kaliszewski <  2008-01-04 18:56:03 
Re: A case for Dual System Cabinets
Sebastian Kaliszewski <  2008-01-04 18:45:47 
Re: A case for Dual System Cabinets
Robert Redelmeier <red  2007-12-27 19:28:32 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-28 03:28:24 
Re: A case for Dual System Cabinets
Robert Redelmeier <red  2007-12-28 15:22:46 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-28 22:15:16 
Re: A case for Dual System Cabinets
chrisv <chrisv@[EMAIL   2007-12-27 10:04:09 
Re: A case for Dual System Cabinets
kony <spam@[EMAIL PROT  2007-12-27 14:23:32 

Post A Reply:
  Go here to Signup

AddThis Feed Button


About - Advertising - Contact - Frequently Asked Questions - Privacy Policy - Terms of Use - Signup

Contact
tan12V112 Sun Nov 23 11:17:11 CST 2008.