From: Charlie Hoffpauir on
On 18 Feb 2010 01:00:22 GMT, John Doe <jdoe(a)usenetlove.invalid> wrote:

>Have you ever actually measured the power that your PC consumes?
>
>I was ready to play along with the "YOU NEED A 1000 WATT POWER
>SUPPLY!!!" people, but apparently that was not the problem here.
>
>The idea that modern systems actually use 500-1000 watts still
>sounds rediculous. The makers increase the power output probably
>just to keep up with the silly idea that a more power hungry
>supply means a faster system and that power requirements are
>increasing as systems get REALLY REALLY FAST!!!
>
>Someday, I need to buy or borrow a power measurement device, the
>kind that clamp onto the power cord without having to touch any
>conductors, then get a real idea of PC power consumption.

Easier than buying a separate device to actually measure the power
used...

http://www.enermax.outervision.com/index.jsp

Just fill in the details of all the components you actually have, and
let their calculator tell you what supply you need.

From: John Doe on
Flasherly <Flasherly(a)live.com> wrote:

> I gots one - @$19 on sales they periodically kick off. Kill-A-
> Watt(tm). Easy way to monitor average costs.

> so I go to a website all about Power Supplies, where the site's
> javascript lets me plugin what hardware I've got, and adds up my
> build's total combined consumption for a determination,
> aroundabout for a 500W PS, which hits the sweet spot;- off and
> over Newegg to skim for efficiency/quality in the reviews. All
> I can say is stick to better-reviewed/regarded units

Why on Earth don't you plug in your PC to the Kill-A-Watt and find
out how much power your PC really uses? I might get one of those.
They are much cheaper than a clamp-on meter, and guessing power
requirements is for the birds.

From an Amazon review of the Kill-A-Watt:
> Core 2 Duo E8600 with 8 GB of ram and 4 hard drives. 100 watts
> all of the time.

HE NEEDS A 750 WATT PS!!!

http://techgage.com/article/intel_core_2_quad_q9450_266ghz/12

Very revealing. Apparently power supply size matters, the same way
penis size matters.




--
Go Chad Hedricks!
Olympic gold medal ice speed skater, also king of inline skating.
From: John Doe on
Charlie Hoffpauir <invalid invalid.com> wrote:

> John Doe <jdoe usenetlove.invalid> wrote:

>>Have you ever actually measured the power that your PC consumes?
>>
>>I was ready to play along with the "YOU NEED A 1000 WATT POWER
>>SUPPLY!!!" people, but apparently that was not the problem here.
>>
>>The idea that modern systems actually use 500-1000 watts still
>>sounds rediculous. The makers increase the power output probably
>>just to keep up with the silly idea that a more power hungry
>>supply means a faster system and that power requirements are
>>increasing as systems get REALLY REALLY FAST!!!
>>
>>Someday, I need to buy or borrow a power measurement device, the
>>kind that clamp onto the power cord without having to touch any
>>conductors, then get a real idea of PC power consumption.
>
> Easier than buying a separate device to actually measure the power
> used...
>
> http://www.enermax.outervision.com/index.jsp
>
> Just fill in the details of all the components you actually have, and
> let their calculator tell you what supply you need.

Have you ever heard of "science"? Neat stuff, try it some time.
--

























> Path: news.astraweb.com!border1.newsrouter.astraweb.com!feed.news.qwest.net!mpls-nntp-08.inet.qwest.net!news.glorb.com!border1.nntp.dca.giganews.com!nntp.giganews.com!npeer03.iad.highwinds-media.com!news.highwinds-media.com!feed-me.highwinds-media.com!spln!extra.newsguy.com!newsp.newsguy.com!news5
> From: Charlie Hoffpauir <invalid invalid.com>
> Newsgroups: alt.comp.hardware.pc-homebuilt
> Subject: Re: Any decent 500-700 watt power supplies no more than $95 US?
> Date: Wed, 17 Feb 2010 21:26:02 -0600
> Organization: As little as possible
> Lines: 25
> Message-ID: <9ncpn55260rr7n17a2t2evq9l3ck57khop 4ax.com>
> References: <4b7b4154$0$15690$c3e8da3 news.astraweb.com> <b9dd072a-3dab-406c-9f83-59d630b2aa1d y33g2000yqb.googlegroups.com> <4b7c7dd4$0$13571$c3e8da3 news.astraweb.com> <deebacc6-b4d5-4366-9482-20c71eecb037 g28g2000yqh.googlegroups.com> <4b7c9126$0$11032$c3e8da3 news.astraweb.com>
> NNTP-Posting-Host: pa419ffaeb6cf26e9e6471839379df018ad6004240fb173b0.newsdawg.com
> Mime-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
> X-Newsreader: Forte Agent 4.2/32.1118
>
From: John Doe on
This is funny IMO.

From below:
> Running Prime95 with all 4 cores loaded (Q9550, in a gaming
> system), power consumption rises to a respectable 245 W.

And the guy does not even wonder why he has a 600 watt PS.




http://www0.epinions.com/review/Intel_Core_2_Quad_Q9550_2_83GHz_Processor_EU80569PJ073N/content_479930715780

Here's the parts list of my rig:

Gigabyte GA-P35-DS3L rev 2.0 motherboard
Intel Core 2 Quad Q9550 G0
CoolerMaster Hyper 212 (heatsink/fan)
Patriot DDR2-800 2 GB (1x 1GB sticks) memory at 5-5-5-16 timings (2.2 volts)
2x Western Digital Raptors at 34 GB each (RAID0)
JMICRO PCI-E controller (for the Raptors in RAID0)
Seagate 500GB SATA drive
EVGA Geforce GTX260 896 MB video card
Seasonic S600 600W PSU
Antec Super Lanboy

They are all wrapped up in Antec's light Lanboy aluminum mid-tower
case. Power consumption was measured with a Kill-A-Watt power
meter.

At the stock 2.83 GHz with SpeedStep (idles at 2 GHz), the entire
rig sucks up 148 W. Not bad for all those components as well as a
quad-core CPU. Running Prime95 with all 4 cores loaded, power
consumption rises to a respectable 245 W.

Repeating the test at 3.5 GHz (Speedstep enabled, but voltage
stepping is set at 1.225v) yielded an idle (2.4 GHz) power
consumption of 165 W. Loaded, the entire rig ate up 280 W. Still
not too bad.

UPDATE: I was able to lower the voltage, so power consumptin also
lowered! At the same speed, I was able to stay stable at 1.183v,
which dropped the loaded power consumption to 260 W, while it
idles around 160W.

Remember that the last time I hit 3.5 GHz with the Q6600 (I had to
disable Speedstep), it ate up 210 W idle, and 300 W loaded! The
lower voltage and temperature really does make a difference!
From: Flasherly on
On Feb 17, 11:27 pm, John Doe <j...(a)usenetlove.invalid> wrote:
> This is funny IMO.


I'm running an older RightMark CPU multiplier, stepping at on or off
the one core I'm using for something around a 3rd less, as it's mostly
off. Your video card series is upwards of 100W draws, though don't
have it's specs per se. Nor have seen what makes the "super" in a
Lanboy (bigger, I'd suspect, though I've seen mine occasionally as low
as $30 on sales -- very nice case in any event). No DVDs, dedicated
raid with a little storage on the side -- actually sounds if
everything could be fitted compactly, as well.

Qualifies your original question, though not as much in a typical
sense a lessor component assembly typifies. Not everyone is running
quadcores, duallies perhaps would be mainstream typical.

Your present Seasonic. Not even going to ask what you paid for that
-- *presently* Seasonic runs upwards of $150+ between 6 and 850W --
nor sure who Newegg's endorsement of JonnyGuru.com qualifies, although
had I a similar setup I'd be in the same boat and looking for the
best.

No biggie - just change the matrics.

Antec has two units -- nothing especially offhand exciting, although
that Corsair CMPSU-650 (top of the reviewer lists by way far ahead)
looks interesting. 5yr warr/variable sensing fan/crossfire(hm)/
*rated* 80% efficiency/brick-heavy/52A 12V rail.

Nor really that much science if it's evidently nailed. Corsair's not
pulling any punches when it comes to quality, all that's left is to
shell out $4 over your limit. Come on... I'd be game, even though it
would be my first Corsair. Last PS I looked hard at was Fortron, and
I was looking only within lowend server-grade units (after getting
over the initial server cost shock). Cost me that when 400W was
considered a highend penis (and I'd dropped near a grand into this
Athlon64 3G I built and am still running).

Sounds like a nice system you have, anyway. Someday I should think
about convincing myself I can do more with four (or even two) than I
do with a single core.