🚀 go-pugleaf

Sorry, the Duron speed is 750 MHz, but that doesn't really affect
the story.
Oliver

In article <1077232439.30270.0@damia.uk.clara.net>, Oliver Walter
<news@owalter.invert> writes
>I've just bought a power consumption meter from Lidl, for �7;
>they're selling it for the next few days.

What, in all stores? I want one :-)

>I was rather surprised that my PC only consumes 140 W including the
>monitor (CD not reading or writing).
>The system is: Daewoo 15" CRT monitor (7 years old, measured 25
>Watts for this), Duron 650 MHz, 2 hard disks, Soltek mobo, 256 MB
>RAM, one floppy drive, cheap Arowana speakers, LiteOn CDRW, no case
>fan, AOpen desktop case

115W exc. monitor seems like a lot for that. What's the video card?


--
John Jordan

The message <XkDl3qTXUVNAFwCf@jaj22.demon.co.uk>
from John Jordan <john@jaj22.demon.co.uk> contains these words:

> In article <1077232439.30270.0@damia.uk.clara.net>, Oliver Walter
> <news@owalter.invert> writes
> >I've just bought a power consumption meter from Lidl, for �7;
> >they're selling it for the next few days.

> What, in all stores? I want one :-)

> >I was rather surprised that my PC only consumes 140 W including the
> >monitor (CD not reading or writing).
> >The system is: Daewoo 15" CRT monitor (7 years old, measured 25
> >Watts for this), Duron 650 MHz, 2 hard disks, Soltek mobo, 256 MB
> >RAM, one floppy drive, cheap Arowana speakers, LiteOn CDRW, no case
> >fan, AOpen desktop case

> 115W exc. monitor seems like a lot for that. What's the video card?

 IME, (using a real wattmeter) 14 and 15 inch monitors take around 50 to
60 watts active (depends on picture content) with a modern 17 inch
taking 70 to 80 watts. It seems that 'Power consumption meter' is
under-reading somewhat.

 I'd expect the base unit to be taking around the 100 watt mark give or
take 10. 150 to 170 watt for the lot seems more likely. Have you tried
it with say a 60 or100 watt tungsten GLS lamp to check the calibration?
What range and  power factor figures does this 'power consumption meter'
cover? Can it handle reactive loads correctly, such as a sunbed
(inductive load)?

--
Regards, John.

 To reply directly, please remove "buttplug" .Mail via the
 "Reply Direct" button and Spam-bots will be rejected.

"Johnny B Good" <jcs.computersbutt@plugzetnet.co.uk> wrote in message
news:2004022001115685168@plugzetnet.co.uk...
> The message <XkDl3qTXUVNAFwCf@jaj22.demon.co.uk>
> from John Jordan <john@jaj22.demon.co.uk> contains these words:
>
> > In article <1077232439.30270.0@damia.uk.clara.net>, Oliver Walter
> > <news@owalter.invert> writes
> > >I've just bought a power consumption meter from Lidl, for �7;
> > >they're selling it for the next few days.
>
> > What, in all stores? I want one :-)
>
> > >I was rather surprised that my PC only consumes 140 W including
the
> > >monitor (CD not reading or writing).
> > >The system is: Daewoo 15" CRT monitor (7 years old, measured 25
> > >Watts for this), Duron 650 MHz, 2 hard disks, Soltek mobo, 256 MB
> > >RAM, one floppy drive, cheap Arowana speakers, LiteOn CDRW, no
case
> > >fan, AOpen desktop case
>
> > 115W exc. monitor seems like a lot for that. What's the video
card?
>
>  IME, (using a real wattmeter) 14 and 15 inch monitors take around
50 to
> 60 watts active (depends on picture content) with a modern 17 inch
> taking 70 to 80 watts. It seems that 'Power consumption meter' is
> under-reading somewhat.
>

25 watts for a 15", 7 year old, CRT monitor, sounds more like standby
mode to me.

jcs.computersbutt@plugzetnet.co.uk wrote:

>  IME, (using a real wattmeter) 14 and 15 inch monitors take around 50 to
> 60 watts active (depends on picture content) with a modern 17 inch
> taking 70 to 80 watts. It seems that 'Power consumption meter' is
> under-reading somewhat.
>
>  I'd expect the base unit to be taking around the 100 watt mark give or
> take 10. 150 to 170 watt for the lot seems more likely. Have you tried
> it with say a 60 or100 watt tungsten GLS lamp to check the calibration?
> What range and  power factor figures does this 'power consumption meter'
> cover? Can it handle reactive loads correctly, such as a sunbed
> (inductive load)?

I have a wattmeter which is probably the same as the one sold by LIDL
(bought for 12 GBP in a Maplin half-price sale). I also have some
"real" wattmeters (Valhalla, VIP).

My guess is that the cheap meters take into account power factor (they
read watts, not volt-amps), but assume sinusoidal waveforms. I've found
the readings to be reasonably accurate for running equipment, but wildly
high for, say, a switched-off ATX motherboard (which does draw a little
power).

(By the way, if anyone has a manual for the Valhalla 2000 wattmeter I'd
be interested to hear.)

In my experience, a typical PC uses around 100W, increasing to 140W if
thinking very hard (CPU torture test). A 15" CRT monitor uses about 70W
when in use. The cheap meter is reasonably accurate in these cases.

HTH,
--
Michael Salem

Some interesting replies.
Yes, I have checked with incandescent bulbs. It showed 66 W for a 60
W bulb and 107 W for a 100 W bulb, so I think it's reasonably
accurate.
The video card is a basic one: Pine 8 MB PCI card, using the SIS
6326 chipset.
The meter also shows the power factor, which was close to one (0.93
IIRC)
I don't think the monitor was in standby mode - the PC had just
booted and was displaying the desktop; I too was surprised at the
low value.

My overall surprise was because I had been used to discussions about
whether a 300 W or 350 W PSU would be enough. It how seems this will
only be a problem if a PC has a powerful video card, 1 or 2 case
fans, a faster CPU than mine (750 MHz Duron) and perhaps a few other
add-ons also.
Can somebody supply a relatively ordinary PC equipment list that
consumes 300 W or more as an example?

(I haven't checked the readings today because that would mean
switching off the PC first, and I only want to do that once per
day.)

Oliver

In article <1077278196.1815.0@echo.uk.clara.net>, Oliver Walter
<news@owalter.invert> writes
>
>My overall surprise was because I had been used to discussions about
>whether a 300 W or 350 W PSU would be enough. It how seems this will
>only be a problem if a PC has a powerful video card, 1 or 2 case
>fans, a faster CPU than mine (750 MHz Duron) and perhaps a few other
>add-ons also.

Case fans are usually ~2W each. HDs are worse - 5+7W active and ~30W
during spin-up.

>Can somebody supply a relatively ordinary PC equipment list that
>consumes 300 W or more as an example?

The problem is that the 300W is split between the 12V and 3.3/5V rails.
I have a (JNC case) 300W PSU here with a 10A 12V line, which means that
it could have trouble with a P4 and a single HD.

With AMD systems the problem is more likely to be with the 3.3/5V line
because the CPU core voltage is derived from that. High-end video cards
are all 50W+, and an FX5950 apparently consumes 95W loaded.

The other problem is that PSUs often can't handle anywhere near their
rated wattages...


--
John Jordan

In article <1077278196.1815.0@echo.uk.clara.net>, Oliver Walter
<news@owalter.invert> writes
>My overall surprise was because I had been used to discussions about
>whether a 300 W or 350 W PSU would be enough. It how seems this will
>only be a problem if a PC has a powerful video card, 1 or 2 case
>fans, a faster CPU than mine (750 MHz Duron) and perhaps a few other
>add-ons also.
>Can somebody supply a relatively ordinary PC equipment list that
>consumes 300 W or more as an example?

There's a lot of "my PSU is bigger than your PSU" bragging rights. Not
exactly in the PSU manufacturers best interests to educate people
otherwise :-)

See thread for good info:
  http://forums.silentpcreview.com/viewtopic.php?t075

cheers,
        Martin Reed <martin@oldbakery.net>

The figures are broadly correct.

A PC spends most of its time idle:
o CPU sits at 60-70% under the Halt instruction
---- so heat dissipation depends on processor leakage (idle wattage)
---- before Prescott this was quiet small 15-20W
---- with Prescott (strained silicon) this has risen to around 40W

o Graphics cards can draw 55W, but only if hammering in 3D
---- anti-aliasing & hot RAM soak up the watts

o RAM wattage isn't insignificant, and is continuous
---- but a typical PC doesn't have the 8-rows of SIMMS

Most PCs draw 110-170W - varying with task obviously.
Running a continuous spreadsheet macro that operates on both
hard drive & computation continually for hours would push that
up to probably 200W as a reasonable limit on the hot P4 CPUs.

A Dual-Xeon, with 4x 15.3k-rpm SCSI RAID 1+0, with (very)
high-end graphics card running a CFD (thermal modelling) or an
Architectural (Rendering) task with continuous CPU will happily
exceed 300W draw. Here you need a 460-550W PSU.

Most home/business *workstation* PCs do not see a high load
average over the period of a day - mainly spikes due to compute,
such as a game or photoshop or a compile or rendering etc. In
contrast a dbase server or row of compute blades do see a near
stable-state continuous high wattage output. Different design mkts.
--
Dorothy Bradbury

The message <c14ke1$1ecgkr$1@ID-8943.news.uni-berlin.de>
from "Notty Pine" <me@privacy.net> contains these words:


> "Johnny B Good" <jcs.computersbutt@plugzetnet.co.uk> wrote in message
> news:2004022001115685168@plugzetnet.co.uk...
> > The message <XkDl3qTXUVNAFwCf@jaj22.demon.co.uk>
> > from John Jordan <john@jaj22.demon.co.uk> contains these words:
> >
> > > In article <1077232439.30270.0@damia.uk.clara.net>, Oliver Walter
> > > <news@owalter.invert> writes
> > > >I've just bought a power consumption meter from Lidl, for �7;
> > > >they're selling it for the next few days.
> >
> > > What, in all stores? I want one :-)
> >
> > > >I was rather surprised that my PC only consumes 140 W including
> the
> > > >monitor (CD not reading or writing).
> > > >The system is: Daewoo 15" CRT monitor (7 years old, measured 25
> > > >Watts for this), Duron 650 MHz, 2 hard disks, Soltek mobo, 256 MB
> > > >RAM, one floppy drive, cheap Arowana speakers, LiteOn CDRW, no
> case
> > > >fan, AOpen desktop case
> >
> > > 115W exc. monitor seems like a lot for that. What's the video
> card?
> >
> >  IME, (using a real wattmeter) 14 and 15 inch monitors take around
> 50 to
> > 60 watts active (depends on picture content) with a modern 17 inch
> > taking 70 to 80 watts. It seems that 'Power consumption meter' is
> > under-reading somewhat.
> >

> 25 watts for a 15", 7 year old, CRT monitor, sounds more like standby
> mode to me.

 That's a possibility I suppose. My ancient 12 inch Microvitec colour
SVGA [1] monitor used to drop down to 50 watts in it's 'Standby' mode
(active consumption measured 80 watts!).

[1] SVGA on account it could do 800 by 600 @ 60Hz NI (flicker not a
serious issue on account of the long persistance phosphors used) or even
1024 by 768 @ 43Hz interlaced (accidentally discovered this mode! :-)

--
Regards, John.

 To reply directly, please remove "buttplug" .Mail via the
 "Reply Direct" button and Spam-bots will be rejected.

"Oliver Walter" <news@owalter.invert:oc.ku> wrote in message
news:1077278196.1815.0@echo.uk.clara.net...
...snip...
> I don't think the monitor was in standby mode - the PC had just
> booted and was displaying the desktop; I too was surprised at the
> low value.
Sorry, the monitor *was* in standby mode when it consumed 25 W. I
had switched it on but not yet the PC. I haven't measured what it
consumes when displaying something, and I probably  won't - that
would require fiddling with the socket strips inside the cabinet
housing the system. My measurements were made at the plug that
supplies the whole system.
BTW, an HP 5p scanner consumes 10 W  when quiescent. (It was off
when the other measurements were made, so they still hold   good.)
Thanks for the other answers John, Martin and Dorothy.

Oliver

"Dorothy Bradbury" <dorothy.bradbury@ntlworld.com> wrote:
>The figures are broadly correct.
>
>A PC spends most of its time idle:
>o CPU sits at 60-70% under the Halt instruction

Noooooo. Must.....run.....distributed.....computing.....
project.....cannot.....waste......cycles..... :)


Tim
--
Love is a travelator.

In article <hrvZb.153$3B5.13@newsfe1-win>, Dorothy Bradbury <dorothy.bra
dbury@ntlworld.com> writes
>
>A PC spends most of its time idle:
>o CPU sits at 60-70% under the Halt instruction
>---- so heat dissipation depends on processor leakage (idle wattage)
>---- before Prescott this was quiet small 15-20W

AFAICT a Northwood uses ~33% of maximum power (prime95) when idle and an
Athlon ~66%. 15-20W idle would need a pretty low-end Celeron. Athlons
don't seem to respond too effectively to Halts.

>---- with Prescott (strained silicon) this has risen to around 40W

Plus a bonus for the overstressed VRM :-)

>o Graphics cards can draw 55W, but only if hammering in 3D
>---- anti-aliasing & hot RAM soak up the watts

Actually most graphics cards have a more Athlon-like idle/loaded ratio,
so even the idle wattages can be pretty huge. The recent nVidia cards
(FX5700+) have frequency/voltage reduction in 2d-only mode (similar
thing to Cool & Quiet), but they use this as an excuse for absolutely
massive loaded wattages - 55W for the FX5700U and 95W for the FX5950.

>o RAM wattage isn't insignificant, and is continuous
>---- but a typical PC doesn't have the 8-rows of SIMMS

About how much per 200MHz SDRAM DIMM then? I'm guessing less than 5W...


--
John Jordan

In <r9C6h+URdtNAFwkt@jaj22.demon.co.uk>,
John Jordan <john@jaj22.demon.co.uk> wrote:

> In article <hrvZb.153$3B5.13@newsfe1-win>, Dorothy Bradbury <dorothy.bra
> dbury@ntlworld.com> writes
>>
>>A PC spends most of its time idle:
>>o CPU sits at 60-70% under the Halt instruction
>>---- so heat dissipation depends on processor leakage (idle wattage)
>>---- before Prescott this was quiet small 15-20W
>
> AFAICT a Northwood uses ~33% of maximum power (prime95) when idle and an
> Athlon ~66%. 15-20W idle would need a pretty low-end Celeron. Athlons
> don't seem to respond too effectively to Halts.

Athlons do if you enable a certain flag. In Linux this can be done on
nearly all chipsets with a one-line shell command. On Windows VCool is
the only thing I've heard of that does this and it only supports VIA and
AMD chipsets with NForce in beta only, and not working last time I tried
it.

It makes at least 15C difference to my idle temperatures.

--
My real address is in the Reply-To and includes the .nospam.
See <http://www.realh.co.uk/contact.html> for more reliable contact addresses.

> Noooooo. Must.....run.....distributed.....computing.....
> project.....cannot.....waste......cycles..... :)

Indeed - a lot of consumer & educational machines.

However using distributed-&-idle CPU power is task dependent:
o Easy - compact work-units, no dependency, no latency (SETI)
o Hard - massively CPU-dependent & latency-dependent (Nuke)
---- comes down to Interconnect performance, Myrinet etc

A lot of tasks do fall into the former category though.
Doable for CFD/FEA/earthquake-analysis/weather-forecasting etc.
--
Dorothy Bradbury

> About how much per 200MHz SDRAM DIMM then? I'm guessing less than 5W...

No, I recall it is 8-12W.

A typical PC can thus see 16-22W on RAM, and with 64-bit we may get
more slots allow more rows of cheaper RAM - so that limit may be exceeded.
--
Dorothy Bradbury

Re RAM consumption...
o 2GB of DDR400 in 4x 512MB is about 35 watts
o Each 512MB is around 5-7 watts

2.5V memory uses a surprising amount of power.
Hence DDR2 shortly which will use 1.8V, so combined with
die shrinkage will allow more DIMM sockets re 64-bit.

Even 35W is not a lot of course - but it DOES all add up.
--
Dorothy Bradbury

Unfortunately, scanning thro some papers I notice...
o DDR2 will scale 400-533-667Mhz
---- nothing remarkable there
o Using IDD7 for peak wattage 667Mhz sees 19W on 512MB
---- that is a theoretical figure somewhat impractical
---- so taking 80% of it gives are figure of nearer 15W

Whatever, for a dual-DIMM arrangement in DDR2 we may find
an atypical configuration dissipating 30W and 25-30oC above
ambient. So heatspreaders may be introduced, altho the BTX
layout using RAM-in-line-with-airflow is beneficial.

When RAM is perpendicular to airflow, you will see a 3-4oC bop
in temperatures on the shadowed DIMM - and that will have an
impact on CPU temperatures if they are "next downstream" for
those who remember the early Dual Athlon 1U rack designs.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan for fans, books & other items
http://homepage.ntlworld.com/dorothy.bradbury/panaflo.htm (Direct)