What's the point of overclocking?

Hi.

Why do people feel the need to overclock their hardware, anyway? I
don't overclock, I settle for the speed the hardware I got is meant to
run at. After all, overclocking means you have to spend _more_ money in
the long run for a given speed, not less, since it shortens the
lifespan and you have to buy replacements/upgrades more often. And what
is the point of "small" overclocks like running a 3GHz chip at 3.1 or
3.2 GHz? Are a few extra frames/second on a game going to really make
that much of a difference? I mean, you can't percieve those type of
changes (100fps vs 110fps is not noticeable.). Is it more of a
"prestige" thing to "wow" your friends?

Overclocker type 1.

Gets at or near the speed of the top chip while paying for the bottom chip.
Chip doesn't burn up.

Overclocker type 2.

Goes for the absolute top speed the chip/motherboard/memory combination can
give him. Generates alot of heat. May lose the chip.

How is #1 possible?

It's the same chip rated for different speeds. The chip that tested out at
the factory as unable to run at the fastest speed at stock voltage becomes
the slower chip. But the overclocker can increase the voltage a little bit
on it and go as fast as the fastest chip, or nearly.

#2 will be aiming to exceed the fastest chip's speed and may need exotic
cooling.

Example of #1:

A Sempron 64 2800+ overclocked 50% will run like an Athlon 64 3500+. A very
noticeable difference. Voltage increase, typically, would be up from 1.4v to
1.45v.

Getting itchy to try it?
--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.

"Ed Light" wrote

A Sempron 64 2800+ overclocked 50% will run like an Athlon 64 3500+. A
very noticeable difference. Voltage increase, typically, would be up from
1.4v to 1.45v.

Note: Those are slightly different chips. The Sempron has less cache memory.
Otherwise they are comparable.

The overclocked Sempron with stock cooler will top out at 45 degrees
centigrade under load, depending on good airflow through the case and no
heat wave outside of it. That cpu is ok up to 60C, though it's not good to
get close to it.

--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.

mike3 wrote:
Hi.

Why do people feel the need to overclock their hardware, anyway? I
don't overclock, I settle for the speed the hardware I got is meant to
run at. After all, overclocking means you have to spend _more_ money in
the long run for a given speed, not less, since it shortens the
lifespan and you have to buy replacements/upgrades more often. And what
is the point of "small" overclocks like running a 3GHz chip at 3.1 or
3.2 GHz? Are a few extra frames/second on a game going to really make
that much of a difference? I mean, you can't percieve those type of
changes (100fps vs 110fps is not noticeable.). Is it more of a
"prestige" thing to "wow" your friends?


"Are a few extra frames/second on a game going to really make that
much of a difference?"

Yes, it does. I cranked my 2.8GHz Northwood to 3.2GHz, and now BF2
is smooth on my machine. There was a slight hesitation before the
overclock was applied (and a rough translation of that, is my
frame rate is below 30 FPS). The trick is, to do the tests, find out
what factors affect the overclock, then crank back a little bit from
your top overclock, to leave some margin for hot summer days.
While I should be able to get 3.5GHz, I only managed 3.2GHz stable.

You see, my machine is not a real gamer. All detail settings in the
game are at their lowest level. My frame rate is pretty poor in
fact. But the overclock made the difference.

So if a person is not too greedy, yes, a little overclock can make
a difference. And with only a small difference in operating
temperature.

There are people who compete for the highest overclock possible, who
own $5000 phase change coolers. You don't see too many of them in
USENET groups. If you want to learn about overclocking, in all its
glory, go here. I always find their research and discoveries
fascinating.

http://www.xtremesystems.org/forums/ (scroll down and look at...)

Air Cooling
Liquid Cooling
T.E.C. Cooling
Chilled Liquid Cooling
Vapor Phase Change Cooling
Dry Ice and Liquid Nitrogen

Are there degradation mechanisms ? Yes. One mechanism is called
electromigration. It is affected by heat, voltage, and frequency.
My processor runs cool, so that part is not an issue. My voltage
is also quite reasonable, and is not even near Vcore_max from the
Intel datasheet. Frequency is also part of the electromigration
effect. My frequency is below the max frequency for that family
of processors (they are all based on the same die after all).
The silicon designers design track width for the max expected
frequencies and currents, so the design rules should be sufficient
up to the highest shipping processor frequency. I'm not too worried
about my modest overclock.

http://en.wikipedia.org/wiki/Electromigration

But there are people who have observed overclocking failure. I've
read of a couple cases with Athlon64, where the processor becomes
unstable, the frequency must be reduced, and eventually the processor
will not even run at stock speed any more. As the geometry of the
processors gets smaller, I expect to see more reports like that.
While you're over on xtremesystems, you can have a look and
see if there are any reports like that for Conroe yet.

There are also "death mechanisms". For example, if you had an
FX55, set Vcore high, and left Vdimm at stock, it seemed
the voltage difference could kill the processor outright. There
is no suggestion in the AMD datasheet, of such a mechanism. There
was another issue like that, with Northwood processors. If you
took them up around Vcore_max, they'd go like a lightbulb (no
response on your next attempt to boot). So, yes, there are some
risks, and that is why you visit places like xtremesystems,
because the people there have already sacrificed some processors,
so they can learn about the limits.

HTH,
Paul

Ed Light wrote:
Overclocker type 1.

Gets at or near the speed of the top chip while paying for the bottom chip.
Chip doesn't burn up.

Overclocker type 2.

Goes for the absolute top speed the chip/motherboard/memory combination can
give him. Generates alot of heat. May lose the chip.

How is #1 possible?

It's the same chip rated for different speeds. The chip that tested out at
the factory as unable to run at the fastest speed at stock voltage becomes
the slower chip. But the overclocker can increase the voltage a little bit
on it and go as fast as the fastest chip, or nearly.

#2 will be aiming to exceed the fastest chip's speed and may need exotic
cooling.

Example of #1:

A Sempron 64 2800+ overclocked 50% will run like an Athlon 64 3500+. A very
noticeable difference. Voltage increase, typically, would be up from 1.4v to
1.45v.

Getting itchy to try it?

No. Two reasons:

1. I can live with the speed I've got now, and

2. I don't want to risk the processor.

But I'm curious about this increasingly common (and risky!) practice.
Things are
designed to run a certain way, and I prefer not to push the design of
something
unless I would really NEED to (like saving the Apollo 13 astronauts
where they
pushed the design of the LEM to turn it into a lifeboat.). A chip rated
at X speed
is designed to run at that speed, and no higher. (Going lower might not
be so
bad, but going higher is where one starts to gamble. That's what
overclocking
is -- gambling.)

--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.

"mike3" wrote in message
oups.com...
Hi.

Why do people feel the need to overclock their hardware, anyway? I
don't overclock, I settle for the speed the hardware I got is meant to
run at. After all, overclocking means you have to spend _more_ money in
the long run for a given speed, not less, since it shortens the
lifespan and you have to buy replacements/upgrades more often. And what
is the point of "small" overclocks like running a 3GHz chip at 3.1 or
3.2 GHz? Are a few extra frames/second on a game going to really make
that much of a difference? I mean, you can't percieve those type of
changes (100fps vs 110fps is not noticeable.). Is it more of a
"prestige" thing to "wow" your friends?


Why do people write an anti-overclocking post on an overclocking newsgroup?
Trolling is what that is usually called. Actually we overclock because we
can. The regulars that post here have been doing it for many years. I
seriously doubt that any of them have actually lowered the lifespan of a
processor by overclocking it. Novice overclockers come here to get
information on how to do it safely and reliably. All the major
manufacturers, even Intel now, have all those functions on the motherboards
for overclocking safely, why not use them? The new Core 2 Duo from Intel
will overclock almost 75% (some claim more). My present processor, a
Prescott based EM64T is overclocked from 3.0ghz to 3.6ghz, a 20% increase
and has been that way for 2yrs without a single problem. I notice a huge
increase in video rendering speed. If you don't want to, don't do it. Nobody
is going to force you to. The normal lifespan of a processor is far over
it's useful lifespan. Are you still using the same processor you were using
even 5-6yrs ago? I doubt it unless you are running a linux box.

Ed

"mike3" wrote

A chip rated at X speed is designed to run at that speed, and no higher.

The chip is generally exactly the same one that is in the faster model of
the same type and cache size, or exactly the same except for cache size. Of
course you're not encouraged to know that by the manufacturer. It could be
one that needs a tiny bit more voltage to go fast (they are all tested) or
it could be they had more that could go fast than they needed for the top
model(s), and some of those were used for the slower model.

This
http://www.newegg.com/Product/Produc...82E16819103732
is the same chip as this
http://www.newegg.com/Product/Produc...82E16819103030
and can easily go as fast.

Ditto, this
http://www.newegg.com/Product/Produc...82E16819115005
is the same chip ,except for less cache, as
http://www.newegg.com/Product/Produc...82E16819115002
and can easily go just as fast -- more than $300 worth of speed for free.

If you can afford the fast chips without flinching, then there is no need to
overclock. Otherwise there is no reason not to use the true potential of the
chips, in a conservative fashion.
--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.

Ed Medlin wrote:
"mike3" wrote in message
oups.com...
Hi.

Why do people feel the need to overclock their hardware, anyway? I
don't overclock, I settle for the speed the hardware I got is meant to
run at. After all, overclocking means you have to spend _more_ money in
the long run for a given speed, not less, since it shortens the
lifespan and you have to buy replacements/upgrades more often. And what
is the point of "small" overclocks like running a 3GHz chip at 3.1 or
3.2 GHz? Are a few extra frames/second on a game going to really make
that much of a difference? I mean, you can't percieve those type of
changes (100fps vs 110fps is not noticeable.). Is it more of a
"prestige" thing to "wow" your friends?


Why do people write an anti-overclocking post on an overclocking newsgroup?
Trolling is what that is usually called. Actually we overclock because we
can. The regulars that post here have been doing it for many years. I
seriously doubt that any of them have actually lowered the lifespan of a
processor by overclocking it. Novice overclockers come here to get
information on how to do it safely and reliably. All the major
manufacturers, even Intel now, have all those functions on the motherboards
for overclocking safely, why not use them? The new Core 2 Duo from Intel
will overclock almost 75% (some claim more). My present processor, a
Prescott based EM64T is overclocked from 3.0ghz to 3.6ghz, a 20% increase
and has been that way for 2yrs without a single problem. I notice a huge
increase in video rendering speed. If you don't want to, don't do it. Nobody
is going to force you to. The normal lifespan of a processor is far over
it's useful lifespan. Are you still using the same processor you were using
even 5-6yrs ago? I doubt it unless you are running a linux box.

Ed

I don't consider it trolling, since it's a curiosity question, and I'm
curious if it
really is worth the risk to overclock. But you seem to suggest that for
"modest" overclocks there isn't as much risk as I might have thought,
if it
is done right...

Ed Light wrote:
"mike3" wrote

A chip rated at X speed is designed to run at that speed, and no higher.

The chip is generally exactly the same one that is in the faster model of
the same type and cache size, or exactly the same except for cache size. Of
course you're not encouraged to know that by the manufacturer. It could be
one that needs a tiny bit more voltage to go fast (they are all tested) or
it could be they had more that could go fast than they needed for the top
model(s), and some of those were used for the slower model.


So then why do they sell them as X speed, anyway, if they are "exactly"
the
same?

This
http://www.newegg.com/Product/Produc...82E16819103732
is the same chip as this
http://www.newegg.com/Product/Produc...82E16819103030
and can easily go as fast.

Ditto, this
http://www.newegg.com/Product/Produc...82E16819115005
is the same chip ,except for less cache, as
http://www.newegg.com/Product/Produc...82E16819115002
and can easily go just as fast -- more than $300 worth of speed for free.

If you can afford the fast chips without flinching, then there is no need to
overclock. Otherwise there is no reason not to use the true potential of the
chips, in a conservative fashion.

But if they are no different, why do they sell them as such?

--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.

"mike3" wrote

If you can afford the fast chips without flinching, then there is no need
to
overclock. Otherwise there is no reason not to use the true potential of
the
chips, in a conservative fashion.

But if they are no different, why do they sell them as such?

Capitalism. They charge more for a chip that's set to a faster speed on its
little block, and less for one that's set to a slower speed. Once they have
it mounted on the block and set for a certain speed, then the total package
is different, and they have all the price ranges covered without having to
actually make a bunch of different chips.

If we had no capitalism and had democratic production and distribution of
goods and services, then the chip would be enabled for its maximum
comfortable speed in every instance.

--
Ed Light

Bring the Troops Home:
http://bringthemhomenow.org
http://antiwar.com

Send spam to the FTC at

Thanks, robots.