On 17/03/11 10:59, Daniel James wrote:
In article , Nix wrote:
... which is thousands of times more than any current motherboard
supports.

Not so. My current server ...

Oops, sorry! I thought I had remembered to qualify that remark to
exclude server boards.

... has a motherboard which, fully populated, supports 384Gb

With what size RAM DIMMs? Can you actually achieve that capacity, today,
at any price? If you have 24 slots (which is not impossible) 24x16GB
DIMMs would do it ... and only cost £10k-£15k.

I can achieve that on my 5 year old server, which has 64 slots. With 8GB
DIMMs it could take 512GB of RAM. I don't know about the cost, but
certainly it falls within the remit of "at any price". At the moment it
has a measly 32GB (which was actually quite a lot 5 years ago) and is
sufficient for our current requirements.


If I change "thousands" to "hundreds" my point still stands -- 256TB is
enough addressing space for today's PCs.

It's plenty, and will be for some time to come. At least to those not
trying to model the Earth's atmosphere, or the stars of the Universe.


There are server-class boxes out there with terabytes of RAM now --
although they are hardly consumer-grade.

Indeed they are not.

I suspect boxes like that will be running several CPUs in separate
address spaces, and so aren't strictly relevant to this discussion ...
but I'd be interested to see the spec of such a machine.

My server is an 8-way Opteron SMP box, so all processors are running in
the same address space, and all able to access the full RAM. Small 4-way
SMP boxes with 8- and 12-core Opterons are not uncommon these days. They
are certainly not in the desktop PC category, but neither are they the
preserve of the HPC community either. The building blocks of HPC maybe,
but single units are affordable by small depts. or even motivated
individuals.

--
Nigel Wade

In article , lid
says...
Or this bloke, who makes his own vacuum tubes...
http://highfields-arc.6te.net/constr.../info/f2fo.htm


I get Error 403 when I try to go there.

Yeah - I'm not sure why, it worked yesterday.

--
Skipweasel - never knowingly understood.

On 17/03/11 10:00, John Stumbles wrote:
How does the PC know what sort of monitor it's driving? I didn't realise,
or think of it, that there must be some way the graphics card can query
the monitor for its vertical and horizontal scan capabilities.

From what I understand, new display port connections actually
communicate between card and monitor, and the card will refuse to drive
things that aren't capable.

We recently got a couple of ATI EyeFinity 6-port graphics cards, but
found that they would only drive 6 DP monitors. Anything else is
considered 'legacy' and the card will only drive a maximum of four. It
doesn't matter whether you port a DP adaptor on or not, the card knows
that it's not a DP connection.

chris gurgled happily, sounding much like they were
saying:

How does the PC know what sort of monitor it's driving? I didn't
realise, or think of it, that there must be some way the graphics card
can query the monitor for its vertical and horizontal scan
capabilities.

From what I understand, new display port connections actually
communicate between card and monitor

Good ol' VGA ports have been doing that for years.

On Wed, 16 Mar 2011 22:05:52 +0000, Tom Anderson wrote:

On Wed, 16 Mar 2011, Nix wrote:

I've never entirely understood what USB hubs are supposed to bring to a
monitor?

You plug your keyboard and mouse into the monitor, then the monitor into
the main case. You put the monitor on your desk, and your case on the
floor. Reduces cable clutter.

Some keyboards have USB hubs in them for the similar reasons, but one
advantage of having a hub in the monitor is that it can be powered from
the monitor and thus run more power-hungry connected devices.

--
Anahata
--/-- http://www.treewind.co.uk
+44 (0)1638 720444

On Thu, 17 Mar 2011 14:03:59 +0000, chris wrote:

On 17/03/11 10:00, John Stumbles wrote:
How does the PC know what sort of monitor it's driving?

From what I understand, new display port connections actually
communicate between card and monitor, and the card will refuse to drive
things that aren't capable.

Look up DDC.

I think you'll find if the card doesn't get a response it assumes a safe
default like VGA 640x480 at 60Hz which will get you a good enough display
to use for debugging and fixing things, and is supported by almost any
monitor.

--
Anahata
--/-- http://www.treewind.co.uk
+44 (0)1638 720444

On Wed, 16 Mar 2011 12:53:37 -0000, Daniel James
wrote:

An 80+ "Bronze" certified PSU (the lowest level of 80+ certification)
has to be 86% efficient at 50% but need not be more than 81% efficient
at 20% load.

Using Antec's own online PSU calculator thingie to rate a PSU for the
spec that Phil posted here I get a power requirement of 171W (I'd
guessed about 150W ... but Antec's figures are generous). Antec say

| The total PSU Wattage this tool recommends will give a general
| idea of the range of continuously available power (not peak power)
| at which you should be looking.

The sensible thing to do is to pick a PSU for which 171W (or 150W, if
you prefer my guess) is around 50% of the rated power -- so for Phil's
system that'd be a PSU in the 300-340W range. A more powerful PSU will
cost more and will waste more power, and so will run hotter which will
likely reduce the lifetime of the PSU.

I'd suggest that (say) a Seasonic 330W PSU would be much better choice,
it's one of the few really well made PSUs that has a sensible power
output.

Thanks for the reccomendations

Phil

Cheers,
Daniel.

On 17 Mar 2011 10:00:57 GMT, John Stumbles
wrote:

So now I have a couple of half-decent machines: the Novatech one (quad
core, 4G, but rather tinny box) and the old Targa (AMD64 single core, 1G,
nice box). But you can never have too many PCs ... maybe I will get
around to making that media server so SWMBO & others can watch stuff on
the TV from it.

I seriously recommend getting that copy of this months Computer Shopper
- it really has all you need to get you up to speed. Its number 279
May!! 2011.

Phil

On 16 Mar 2011, Tom Anderson uttered the following:

On Wed, 16 Mar 2011, Nix wrote:
I've never entirely understood what USB hubs are supposed to bring to a monitor?

You plug your keyboard and mouse into the monitor, then the monitor
into the main case. You put the monitor on your desk, and your case on
the floor. Reduces cable clutter.

Ah, right, I see. Something I've been separated from on account of
having a keyboard that lasts forever and uses a PS/2 connector

The ability to build in tinny speakers and SD card readers? (The latter in particular seems deeply arbitrary.)

I find the whole idea of SD card readers pretty baffling, to be
honest. If a device stores data on an SD card, it almost unfailingly
also has a USB port, and it's far easier to plug the device in via
that than it is to fiddle about with tiny little bits of plastic.

A lot of cameras have really really crappy firmware. About 1/5th of the
time, My Canon Digital IXUS (PowerShot IV under a new name) stalls
forever when libgphoto tries to pull photos off it. The SD card reader
always works.

--
NULL && (void)

On 17 Mar 2011, Daniel James outgrape:

In article , Nix wrote:
... which is thousands of times more than any current motherboard
supports.

Not so. My current server ...

Oops, sorry! I thought I had remembered to qualify that remark to
exclude server boards.

True. Desktops are much more limited (my current one has 12Gb and
it turns out not to supply enough power to run that much RAM at
its rated top speed... a sign of untrodden snow.)

... has a motherboard which, fully populated, supports 384Gb

With what size RAM DIMMs? Can you actually achieve that capacity, today,
at any price? If you have 24 slots (which is not impossible) 24x16GB
DIMMs would do it ... and only cost £10k-£15k.

I have no idea I'm fairly sure you couldn't when I bought it, back in
2009, but I suppose the point of this sort of machine is to be crazy
extensible and somewhat futureproofed.

If I change "thousands" to "hundreds" my point still stands -- 256TB is
enough addressing space for today's PCs.

True. I was being pedantic

[Tb RAM]
I suspect boxes like that will be running several CPUs in separate
address spaces, and so aren't strictly relevant to this discussion ...
but I'd be interested to see the spec of such a machine.

SGI-as-was sold quite a lot of them, many very very multiprocessor
indeed: that's why the Linux kernel is known to scale to thousands of
CPUs.

--
NULL && (void)

On Fri, 18 Mar 2011 00:25:53 +0000, Nix wrote:

On 16 Mar 2011, Tom Anderson uttered the following:

On Wed, 16 Mar 2011, Nix wrote:
I've never entirely understood what USB hubs are supposed to bring to
a monitor?

You plug your keyboard and mouse into the monitor, then the monitor
into the main case. You put the monitor on your desk, and your case on
the floor. Reduces cable clutter.

Ah, right, I see. Something I've been separated from on account of
having a keyboard that lasts forever and uses a PS/2 connector

Model M?



--
Use the BIG mirror service in the UK:
http://www.mirrorservice.org

*lightning protection* - a w_tom conductor

On 17 Mar 2011, Nigel Wade stated:

My server is an 8-way Opteron SMP box, so all processors are running in
the same address space, and all able to access the full RAM. Small 4-way
SMP boxes with 8- and 12-core Opterons are not uncommon these days. They

Pedant point: Intel multisocket systems are not SMP these days, and I'm
fairly sure the same is true of AMD as well. They're NUMA, with some of
the RAM accessible only through some of the scokets.

--
NULL && (void)

On 16 Mar 2011, Tony Houghton outgrape:

In ,
Nix wrote:

[ATI video cards]

... and has reasonably good support in the open Linux drivers. (The
only thing that isn't supported and never will be is hardware video
decoding.)

But hopefully there will be OpenCL video decoders and processors before
long?

As long as they can do that without letting the bloody DRM cat out of
the bag. I think they want it to work as a video card first.

--
NULL && (void)

On Fri, 18 Mar 2011 00:28:28 +0000, Nix wrote:

SGI-as-was sold quite a lot of them, many very very multiprocessor
indeed: that's why the Linux kernel is known to scale to thousands of
CPUs.

How do they fit all the penguins on the screen when it's booting?

;-)

--
John Stumbles

On Thu, 17 Mar 2011 22:32:52 +0000
Phil Addison wrote:

On 17 Mar 2011 10:00:57 GMT, John Stumbles
wrote:

So now I have a couple of half-decent machines: the Novatech one
(quad core, 4G, but rather tinny box) and the old Targa (AMD64
single core, 1G, nice box). But you can never have too many PCs ...
maybe I will get around to making that media server so SWMBO &
others can watch stuff on the TV from it.

I seriously recommend getting that copy of this months Computer
Shopper
- it really has all you need to get you up to speed. Its number 279
May!! 2011.

Phil

Yep, I got it. I won't be building my own, thank you.
--
Davey.

On Thu, 17 Mar 2011 10:25:12 -0500, anahata
wrote:

On Thu, 17 Mar 2011 14:03:59 +0000, chris wrote:

On 17/03/11 10:00, John Stumbles wrote:
How does the PC know what sort of monitor it's driving?

From what I understand, new display port connections actually
communicate between card and monitor, and the card will refuse to drive
things that aren't capable.

Look up DDC.

I think you'll find if the card doesn't get a response it assumes a safe
default like VGA 640x480 at 60Hz which will get you a good enough display
to use for debugging and fixing things, and is supported by almost any
monitor.

You'd hope. On my Ubuntu 8.something it assumes my monitor is a
widescreen running at high resolution, when connected via the KVM. :-(

Easily fixed by tweaking the config though. I don't know why they
can't just connect through pin 12.
--
(\__/) M.
(='.'=) Due to the amount of spam posted via googlegroups and
(")_(") their inaction to the problem. I am blocking some articles
posted from there. If you wish your postings to be seen by
everyone you will need use a different method of posting.

On 17/03/11 15:25, anahata wrote:
On Thu, 17 Mar 2011 14:03:59 +0000, chris wrote:

On 17/03/11 10:00, John Stumbles wrote:
How does the PC know what sort of monitor it's driving?

From what I understand, new display port connections actually
communicate between card and monitor, and the card will refuse to drive
things that aren't capable.

Look up DDC.

I think you'll find if the card doesn't get a response it assumes a safe
default like VGA 640x480 at 60Hz which will get you a good enough display
to use for debugging and fixing things, and is supported by almost any
monitor.

This is different. The card won't drive more than 4 non-Display Port
monitors at all. I don't fully understand why, but it seems that DP
connections are less arduous than DVI/VGA connections for cards to drive.

On 18/03/11 00:30, Nix wrote:
On 17 Mar 2011, Nigel Wade stated:

My server is an 8-way Opteron SMP box, so all processors are running in
the same address space, and all able to access the full RAM. Small 4-way
SMP boxes with 8- and 12-core Opterons are not uncommon these days. They

Pedant point: Intel multisocket systems are not SMP these days, and I'm
fairly sure the same is true of AMD as well. They're NUMA, with some of
the RAM accessible only through some of the scokets.


Yes, it's NUMA rather than SMP. But all the RAM is accessible by all
CPUs. Each CPU can access "local" RAM fastest, with slower access to RAM
in other "nodes". But each CPU has access to all the RAM, should it
require it.

I tend to think in terms of SMP vs. clusters. Where SMP is lots of
processors and RAM in one box, where any process can run on any CPU and
access all the memory. Whereas a cluster is a maze of twisty little
boxes, all alike.

--
Nigel Wade

In article , Nigel Wade wrote:
I can achieve that on my 5 year old server, which has 64 slots. With
8GB DIMMs it could take 512GB of RAM.

That is a fruitbat of a lot of slots ... but (as I noted) I had intended
to exclude server boards.

certainly it falls within the remit of "at any price".

Yes, it does. My "at any price" remark was meant to stop people saying
"Ah, but if 256GB DIMMs were available then I could ...". AFAIK 16GB is
as big a DIMM as one can currently buy (and they're around £600 each
from Crucial) ... and not all memory controllers support them.

It's plenty, and will be for some time to come. At least to those not
trying to model the Earth's atmosphere, or the stars of the Universe.

I'd like enough memory to have one bit to model the state of every
fundamental particle in the universe ... but I haven't got a universe
large enough to put it in ... smile

I suspect boxes like that will be running several CPUs in separate
address spaces, and so aren't strictly relevant to this discussion
... but I'd be interested to see the spec of such a machine.

My server is an 8-way Opteron SMP box, so all processors are running
in the same address space, and all able to access the full RAM. Small
4-way SMP boxes with 8- and 12-core Opterons are not uncommon these
days. They are certainly not in the desktop PC category, but neither
are they the preserve of the HPC community either.

Indeed, but I was replying to Nix's remark about

... server-class boxes out there with terabytes of RAM now ...

... and while it's not impossible to build an SMP Opteron system with
2TB+ (128+ slots each holding a 16GB DIMM) I don't know of any hardware
than can do it, and I suspect there is very little, if any.

There certainly are servers with more total RAM in multiple address
spaces, though. As the number of processors/cores accessing RAM
increases the memory bus becomes a bottleneck, and you get more
efficient memory usage in non-SMP architectures with multiple memory
buses. As these new architectures gain importance I'd think it's
becoming less and less likely that anyone will build a multi-terabyte
SMP system.

Cheers,
Daniel.

In article , Phil Addison
wrote:
I'd suggest that (say) a Seasonic 330W PSU would be much better
choice, it's one of the few really well made PSUs that has a
sensible power output.

Thanks for the reccomendations

FWIW I have a Seasonic 330W PSU in a system with an (older model, socket
AM2) Athlon X2 4200+ EE in an Asus M2NPV-VM uATX motherboard (with
onboard nVidia graphics) with 6GB or RAM. The whole system with a
vanilla DVD drive and a single SATA HDD draws a little over 65W at the
mains when idle. 330W is really too much ...

Cheers,
Daniel.

On Mar 17, 10:59*am, Daniel James wrote:

Today it's not a big deal ... but if you have a need for a USB hub at
all it saves a bit of space to have it built into the monitor rather

Except the totally crap way Dell do it. Turn off the monitor and the
built-in hub is useless until you force a re-enumeration by
disconnecting and reconnecting the USB from the PC.

MBQ

We were somewhere around Barstow, on the edge of the desert, when the
drugs began to take hold. I remember Bob Eager
saying something like:

Ah, right, I see. Something I've been separated from on account of
having a keyboard that lasts forever and uses a PS/2 connector

Model M?

waves

In article , Daniel James
scribeth thus
In article , Phil Addison
wrote:
I'd suggest that (say) a Seasonic 330W PSU would be much better
choice, it's one of the few really well made PSUs that has a
sensible power output.

Thanks for the reccomendations

FWIW I have a Seasonic 330W PSU in a system with an (older model, socket
AM2) Athlon X2 4200+ EE in an Asus M2NPV-VM uATX motherboard (with
onboard nVidia graphics) with 6GB or RAM. The whole system with a
vanilla DVD drive and a single SATA HDD draws a little over 65W at the
mains when idle. 330W is really too much ...

Cheers,
Daniel.



Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think makers
do them in the Hundreds of watts range?..

--
Tony Sayer

On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

In article , Daniel James
scribeth thus
In article , Phil
Addison wrote:
I'd suggest that (say) a Seasonic 330W PSU would be much better
choice, it's one of the few really well made PSUs that has a
sensible power output.

Thanks for the reccomendations

FWIW I have a Seasonic 330W PSU in a system with an (older model,
socket AM2) Athlon X2 4200+ EE in an Asus M2NPV-VM uATX motherboard
(with onboard nVidia graphics) with 6GB or RAM. The whole system
with a vanilla DVD drive and a single SATA HDD draws a little over
65W at the mains when idle. 330W is really too much ...

Cheers,
Daniel.



Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?
--
Davey.

In article , Davey
scribeth thus
On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

In article , Daniel James
scribeth thus
In article , Phil
Addison wrote:
I'd suggest that (say) a Seasonic 330W PSU would be much better
choice, it's one of the few really well made PSUs that has a
sensible power output.

Thanks for the reccomendations

FWIW I have a Seasonic 330W PSU in a system with an (older model,
socket AM2) Athlon X2 4200+ EE in an Asus M2NPV-VM uATX motherboard
(with onboard nVidia graphics) with 6GB or RAM. The whole system
with a vanilla DVD drive and a single SATA HDD draws a little over
65W at the mains when idle. 330W is really too much ...

Cheers,
Daniel.



Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?

Yes, now what's the watts when there're not idle;?..
--
Tony Sayer

In article ,
tony sayer wrote:
In article , Davey
scribeth thus
On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?

Yes, now what's the watts when there're not idle;?..

The worst case will probably be at power-on, when everything is active
at once, including spinning up disc drives, CD/DVD drive, etc. I CBA
to find out my power meter but I think I get about 250W drawn from the
mains at startup for a system which draws 60W when idle - and that's
without a snazzy graphics card etc. My PSU is probably over-specced at
450W but IME the lower powered ones die sooner, as you said !

This point about PSU's being very inefficient at low outputs is one I
didn't know though. Hmmm.

Nick
--
Serendipity: http://www.leverton.org/blosxom (last update 29th March 2010)
"The Internet, a sort of ersatz counterfeit of real life"
-- Janet Street-Porter, BBC2, 19th March 1996

On Fri, 18 Mar 2011 22:48:53 +0000 (UTC)
Nick Leverton wrote:

In article ,
tony sayer wrote:
In article , Davey
scribeth thus
On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

Won't harm 'tho will it. I have noticed that overrated PSU's seem
to last that bit longer .. in fact a lot longer;!..
little bit

Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?

Yes, now what's the watts when there're not idle;?..

The worst case will probably be at power-on, when everything is active
at once, including spinning up disc drives, CD/DVD drive, etc. I CBA
to find out my power meter but I think I get about 250W drawn from the
mains at startup for a system which draws 60W when idle - and that's
without a snazzy graphics card etc. My PSU is probably over-specced
at 450W but IME the lower powered ones die sooner, as you said !

This point about PSU's being very inefficient at low outputs is one I
didn't know though. Hmmm.

Nick

They may be inefficient, but:
1. Do they need to have the full power output at power-on, as
described, in which case, they are needed;
2. Is it worth worrying about, if they last a lot longer;
3. Does it matter? Ok, we all want to be efficient, but if it only
draws a little more current than a more efficient PSU, so what? You
could turn a light out to compensate. Overall, the higher power one
might be the best buy, if it doesn't need to be replaced.

Just thinking out loud. It's a little like a guy I knew who had a Lotus
Esprite, he was worried about whether to buy the sticky-but-short-lived
tyres, or the not-so-sticky-but-longer-lived ones. Since he only drove
it at weekends, he would probably never get through the first set. QED.
--
Davey.

On 18 Mar 2011, Bob Eager spake thusly:

On Fri, 18 Mar 2011 00:25:53 +0000, Nix wrote:
Ah, right, I see. Something I've been separated from on account of
having a keyboard that lasts forever and uses a PS/2 connector

Model M?

No, something much rarer: a Maltron. I tried using a Model M a few years
ago and it was even worse than a normal keyboard: normal keyboards are
like hot wires in the back of my hand, the Model M was like running hot
lead in my bones.

So, not a Model M fan here. Also they're too noisy: if anyone in my
office environment got something that loud I'd have to kill him.
Maltrons use Cherry MX Black keyswitches, which are *quiet*.

--
NULL && (void)

On 18 Mar 2011, Nigel Wade uttered the following:

On 18/03/11 00:30, Nix wrote:
On 17 Mar 2011, Nigel Wade stated:

My server is an 8-way Opteron SMP box, so all processors are running in
the same address space, and all able to access the full RAM. Small 4-way
SMP boxes with 8- and 12-core Opterons are not uncommon these days. They

Pedant point: Intel multisocket systems are not SMP these days, and I'm
fairly sure the same is true of AMD as well. They're NUMA, with some of
the RAM accessible only through some of the scokets.

Yes, it's NUMA rather than SMP. But all the RAM is accessible by all
CPUs. Each CPU can access "local" RAM fastest, with slower access to RAM
in other "nodes". But each CPU has access to all the RAM, should it
require it.

Yeah, sorry, I picked the wrong word. I meant *directly* accessible:
the other RAM is indirectly accessible.

--
NULL && (void)

In article , Nick Leverton
scribeth thus
In article ,
tony sayer wrote:
In article , Davey
scribeth thus
On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?

Yes, now what's the watts when there're not idle;?..

The worst case will probably be at power-on, when everything is active
at once, including spinning up disc drives, CD/DVD drive, etc. I CBA
to find out my power meter but I think I get about 250W drawn from the
mains at startup for a system which draws 60W when idle - and that's
without a snazzy graphics card etc. My PSU is probably over-specced at
450W but IME the lower powered ones die sooner, as you said !


Which is really the point. Whatever the system is doing on "tickover" is
not relevant to what its doing under sail when the watts will really be
needed...

This point about PSU's being very inefficient at low outputs is one I
didn't know though. Hmmm.

Dunno really most of the switch mode supplies we use here and there very
rarely seem to get hot at all up to their rated loads, unlike the linear
versions...


Nick

--
Tony Sayer

Just thinking out loud. It's a little like a guy I knew who had a Lotus
Esprite,

Oi!, it's Esprit.. Had to sell mine, couldn't fit in it with the advancing
years and waist measurement let alone the height...

he was worried about whether to buy the sticky-but-short-lived
tyres, or the not-so-sticky-but-longer-lived ones. Since he only drove
it at weekends, he would probably never get through the first set. QED.


Dunno .. If you used them like you could you'd get thru them..

Haven't seen one of these of any variant around on the roads for quite some
time now. I suppose people are preserving them and don't want to risk them
being damaged as they didn't hold together all that well in crashes..

Let alone the olde Colin Chapman adage that,

"If it hangs together for more than the one race .. we've built it too well"

--
Tony Sayer

On Sun, 20 Mar 2011 15:32:45 +0000
tony sayer wrote:


Just thinking out loud. It's a little like a guy I knew who had a
Lotus Esprite,

Oi!, it's Esprit.. Had to sell mine, couldn't fit in it with the
advancing years and waist measurement let alone the height...

Yep, sorry. Got mixed up with my old Sprite. Performance was
different, though.

he was worried about whether to buy the sticky-but-short-lived
tyres, or the not-so-sticky-but-longer-lived ones. Since he only
drove it at weekends, he would probably never get through the first
set. QED.


Dunno .. If you used them like you could you'd get thru them..

Haven't seen one of these of any variant around on the roads for
quite some time now. I suppose people are preserving them and don't
want to risk them being damaged as they didn't hold together all that
well in crashes..

Let alone the olde Colin Chapman adage that,

"If it hangs together for more than the one race .. we've built it
too well"


The one I referred to was in fact in Tennessee, and was definitely a
second car, for occasional use. The roads around the Saturn car factory
grounds were unlimited, for plant employees like the owner, and
unoccupied on Sunday afternoons.
Fun!
--
Davey.

In article , Tony sayer wrote:
FWIW I have a Seasonic 330W PSU in a system with ...
[snip]
330W is really too much ...

Won't harm 'tho will it.

A lower-rated PSU might have been a little cheaper, had one been
available, and the average current drawn by the system would have been
closer to the maximum efficiency point of the PSU ...

All in all it might have saved a couple of beers over the lifetime of
the PC.

Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..

It'll draw maybe 110W when working hard. I haven't pushed it hard with
the meter connected ...

Several reasons, though: Mostly because people don't know how much
power a PC is actually going to draw and assume "bigger is better";
partly because people observe that when they buy cheap trashy PSUs they
need to get one with a ridiculously high rating just to get stable
voltages and low ripple; and partly because some people put a couple of
200W GPUs in their systems while I (like the OP) am using onboard
graphics that draws maybe 20W on a bad day.

Cheers,
Daniel.

On 21/03/11 12:58, Daniel James wrote:
In , Tony sayer wrote:
FWIW I have a Seasonic 330W PSU in a system with ...
[snip]
330W is really too much ...

Won't harm 'tho will it.

A lower-rated PSU might have been a little cheaper, had one been
available, and the average current drawn by the system would have been
closer to the maximum efficiency point of the PSU ...

All in all it might have saved a couple of beers over the lifetime of
the PC.

Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..

It'll draw maybe 110W when working hard. I haven't pushed it hard with
the meter connected ...

Several reasons, though: Mostly because people don't know how much
power a PC is actually going to draw and assume "bigger is better";
partly because people observe that when they buy cheap trashy PSUs they
need to get one with a ridiculously high rating just to get stable
voltages and low ripple; and partly because some people put a couple of
200W GPUs in their systems while I (like the OP) am using onboard
graphics that draws maybe 20W on a bad day.

Isn't it because the wattage is the total over all the 'rails', but the
individual limits on each rail are much lower. Good PSUs will deliver
decent wattages over all the rails, whereas poor ones will skew the
individual wattages to give the 'best' overall wattage.

At least that's how I understood it, last time I looked at buying PSUs.

On 18/03/11 22:48, Nick Leverton wrote:
In article ,
tony sayer wrote:
In article , Davey
scribeth thus
On Fri, 18 Mar 2011 18:50:29 +0000
tony sayer wrote:

Won't harm 'tho will it. I have noticed that overrated PSU's seem to
last that bit longer .. in fact a lot longer;!..


Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..


"65 watts When idle"?

Yes, now what's the watts when there're not idle;?..

The worst case will probably be at power-on, when everything is active
at once, including spinning up disc drives, CD/DVD drive, etc. I CBA
to find out my power meter but I think I get about 250W drawn from the
mains at startup for a system which draws 60W when idle - and that's
without a snazzy graphics card etc. My PSU is probably over-specced at
450W but IME the lower powered ones die sooner, as you said !

Maybe because the low powered ones that you see failing are the very
cheap ones.

There are good quality low wattage PSUs, just as for higher wattage.
There are, however, a whole lot more cheap, low powered, PSUs.
Especially in systems sold by box shifters. A cheap PSU is a simple way
of shaving a few pounds off the total cost.

--
Nigel Wade

In article , Daniel James
scribeth thus
In article , Tony sayer wrote:
FWIW I have a Seasonic 330W PSU in a system with ...
[snip]
330W is really too much ...

Won't harm 'tho will it.

A lower-rated PSU might have been a little cheaper, had one been
available, and the average current drawn by the system would have been
closer to the maximum efficiency point of the PSU ...

All in all it might have saved a couple of beers over the lifetime of
the PC.

Q.. If your PC seems to be drawing just 65 watts why do you think
makers do them in the Hundreds of watts range?..

It'll draw maybe 110W when working hard. I haven't pushed it hard with
the meter connected ...

Several reasons, though: Mostly because people don't know how much
power a PC is actually going to draw and assume "bigger is better";
partly because people observe that when they buy cheap trashy PSUs they
need to get one with a ridiculously high rating just to get stable
voltages and low ripple; and partly because some people put a couple of
200W GPUs in their systems while I (like the OP) am using onboard
graphics that draws maybe 20W on a bad day.

Cheers,
Daniel.




Sometimes people do have more than the one drive, and other stuff still
not a lot of difference in price. I paid a bit more to get a quiet
one..
--
Tony Sayer

Meanwhile, at the uk.comp.os.linux Job Justification Hearings, chris chose
the tried and tested strategy of:

Isn't it because the wattage is the total over all the 'rails', but the
individual limits on each rail are much lower. Good PSUs will deliver
decent wattages over all the rails, whereas poor ones will skew the
individual wattages to give the 'best' overall wattage.

These http://www.playtool.com/pages/psucompat/compat.html eight pages are
everything you ever wanted to know about PSUs, but were afraid to ask. The
page with photos of connectors on is especially useful.

--
http://ale.cx/ (AIM:troffasky) )
19:50:03 up 1 day, 22:08, 6 users, load average: 0.16, 0.07, 0.10
"I am utterly appalled at how I have been treated like a criminal"
-- Andrew Crossley, ACS:Law, 13 August 2010

We were somewhere around Barstow, on the edge of the desert, when the
drugs began to take hold. I remember John Stumbles
saying something like:

On that basis I'm thinking this one from Novatech might be OK
http://www.novatech.co.uk/novatech/p...BB-6404GA.html
I note that it can do 8G RAM but only comes with 2, as 2*2G modules using
up both slots, so if I wanted more I'd have to see if they'd do it at
time of sale otherwise I'd be throwing away the existing 2G modules to
install 4G ones.

Many thanks for this post. It spurred me into finally updating my main
box, which was struggling at bit as it was 5 years old.
Now it's sporting a Phenom II 840 quad core (1), w.4GB ram and a
reasonably future-proof m'bd, for pretty much the same price as novatech
were quoting. I decided to hang on to the good bits of my system - the
PSU and case are excellent and have been customised over the years, so I
was reluctant to let them go and truth to tell, I wasn't too impressed
with the cheap PSU/case of Novatech's offering.

Woohoo. Now I can play HD video with ease.

(1) I know it's really an Athlon, but it does the job and a six-core can
be fitted in a couple of years.