Is overclocking genuinely useful?

[Doc]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo 2.33 gig
processor. The CPU is just under specs for some newer video games.
Wondering about overclocking to move it into a higher realm of
performance to avoid popping for another CPU.

Apparently with this mobo it can be done but I wonder if it's really a
practical solution. Has anyone here done overclocking successfully and
found the computer to be as reliable as it is without o/c'ing?

 

[John Doe]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo
2.33 gig processor. The CPU is just under specs for some newer
video games. Wondering about overclocking to move it into a
higher realm of performance to avoid popping for another CPU.

That depends on what "higher realm" means.

Apparently with this mobo it can be done but I wonder if it's
really a practical solution. Has anyone here done overclocking
successfully and found the computer to be as reliable as it is
without o/c'ing?

There are forums about overclocking. Probably mostly on the web.

Generally... Overclocking means more heat and potentially more
errors.

 

[Paul]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo 2.33 gig
processor. The CPU is just under specs for some newer video games.
Wondering about overclocking to move it into a higher realm of
performance to avoid popping for another CPU.

Apparently with this mobo it can be done but I wonder if it's really a
practical solution. Has anyone here done overclocking successfully and
found the computer to be as reliable as it is without o/c'ing?

It worked for me. I had a game, with a stuttering problem, and
using about a 10% overclock, was enough to fix it.

For many other purposes, that small an overclock is
completely invisible. But that one game, it made a difference.

When you overclock, you approach the instability point (with my method,
a problem pops up about every ten minutes). Then, you adjust the
overclock downwards a notch or two, and do an overnight test for
stability. And that value of overclock, is your "everyday overclock"
value. You leave a little room, for aging.

Even having done that, eventually, the processor was no longer
stable, and I had to turn it down another notch. That computer
is no longer used, and when it comes out of storage for a run,
I just leave it at stock now. That processor was never that
good an overclocker to begin with - and that's the nature
of overclocking. Not every processor is a winner, with gobs
of headroom to harvest. Some barely overclock at all. And
it can also be the motherboard which is the limiting factor.
For example, an experimenter in Japan, using my model of motherboard,
was able to squeeze out another 500MHz of overclocking, by improving
the bypass capacitors (ceramics) used on the motherboard. There are all
sorts of tricks... I can't get the components here, that he
was using. Chances are, you could only get them in China or
Japan.

Paul

 

[SteveH]

Paul presented the following explanation :

It worked for me. I had a game, with a stuttering problem, and
using about a 10% overclock, was enough to fix it.

For many other purposes, that small an overclock is
completely invisible. But that one game, it made a difference.

When you overclock, you approach the instability point (with my method,
a problem pops up about every ten minutes). Then, you adjust the
overclock downwards a notch or two, and do an overnight test for
stability. And that value of overclock, is your "everyday overclock"
value. You leave a little room, for aging.

Even having done that, eventually, the processor was no longer
stable, and I had to turn it down another notch. That computer
is no longer used, and when it comes out of storage for a run,
I just leave it at stock now. That processor was never that
good an overclocker to begin with - and that's the nature
of overclocking. Not every processor is a winner, with gobs
of headroom to harvest. Some barely overclock at all. And
it can also be the motherboard which is the limiting factor.
For example, an experimenter in Japan, using my model of motherboard,
was able to squeeze out another 500MHz of overclocking, by improving
the bypass capacitors (ceramics) used on the motherboard. There are all
sorts of tricks... I can't get the components here, that he
was using. Chances are, you could only get them in China or
Japan.

Paul

Then again, he might just need a new video card.

 

[Darklight]

That depends on what "higher realm" means.


There are forums about overclocking. Probably mostly on the web.

Generally... Overclocking means more heat and potentially more
errors.

A decent heat sink is what you need to ensure the life of your cpu.
When it comes to over clocking a cpu. Heat is your enemy!

example i have a amd fx-8120 overclocked to 3921MHz the standard speed
is 3100MHz.

with stock stock cpu cooler and running prime95 i got scared and stopped
prime95 after cpu temp got to 67c that was within 5mins. .

using after market cpu cooler i run prime95 and temp stays stable at 58c.

and yes with games you do notice the difference. battlefield 3 at stock
speeds seems slow after overclocking it flys.

 

[Paul]

Then again, he might just need a new video card.

It's a balance thing. You need some of both. Some games
are CPU intensive, others GPU intensive. And sometimes, when
you see stutter, it's because the main game thread is dropping
frames. And a CPU core needs to run a bit faster to help the
main thread.

I can also screw up my game, by turning on anti-aliasing, and
then the frame rate plummets. If I want anti-aliasing, that
indeed would take a better video card. One with "legs"
(more memory bandwidth on the video card).

Paul

 

[Flasherly]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo 2.33 gig
processor. The CPU is just under specs for some newer video games.
Wondering about overclocking to move it into a higher realm of
performance to avoid popping for another CPU.

Apparently with this mobo it can be done but I wonder if it's really a
practical solution. Has anyone here done overclocking successfully and
found the computer to be as reliable as it is without o/c'ing?

The practical solution is it would be sold for more and run better and
faster. And that's why cores are locked off from defective batches
and sold for less for fewer cores with an upper-limit throttle on the
speed multiplier. But I'm not as practical as most about getting from
one stoplight to another, at 100mph, in time to leave a trail of burnt
rubber from applying the brakes, either. Maybe if I skipped 1st and
3rd. Being a manual, it's kind of expected, too.

 

[Jon Danniken]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo 2.33 gig
processor. The CPU is just under specs for some newer video games.
Wondering about overclocking to move it into a higher realm of
performance to avoid popping for another CPU.

Apparently with this mobo it can be done but I wonder if it's really a
practical solution. Has anyone here done overclocking successfully and
found the computer to be as reliable as it is without o/c'ing?

The biggest part of overclocking is keeping the computer reliable, this
is why any overclock includes burn-in and stability testing (using
applications such as memtest86 and prime95/orthos).

Basically you increase the clock speed until you get errors, then you
increase the voltage until the errors go away. More voltage means more
heat, so at this point you may need to reconsider your thermal solution.
More voltage also means less component life, but how much depends
upon your components, and on your overclock.

If your a geeky kind of guy, it can be a fun exercise to see just what
your kit is capable of, and run a benchmark where you can actually get a
number that quantifies your performance.

Jon

 

[larry moe 'n curly]

I currently have a Gigabyte GA-P35-DS3L mobo with a Core2 Duo 2.33 gig
processor. The CPU is just under specs for some newer video games.
Wondering about overclocking to move it into a higher realm of
performance to avoid popping for another CPU.

Apparently with this mobo it can be done but I wonder if it's really a
practical solution. Has anyone here done overclocking successfully and
found the computer to be as reliable as it is without o/c'ing?

Check overclocking websites like HardOCP.com , Overclockers.com , and
Overclockers.com.au

XbitLabs did a review of the motherboard and found it was a good
overclocker:

http://www.xbitlabs.com/articles/mainboards/display/ga-p35-ds3l_7.html#sect0

Overclocking your 65W E6540 or E6550 CPU shouldn't come anywhere close
to straining this motherboard, which is rated for 135W CPUs. Some 95W
motherboards have blown out capacitors or MOSFETs while running
overclocked 65W CPUs.

But reliability is never as good with overclocking as without it, so
test thoroughly, at a case temperature slightly hotter than the usual
worst-case (turn off fans to raise temp). Also if you're using
marginal memory, i.e., anything with heatsinks on it or rated for more
than 1.8V (DDR2) or 1.5V (DDR3), its chips may already be overclocked
(APHnetworks.com showed some memory advertised for 2133 MHz being made
from 1333 MHz chips). Heatsinks on memory are actually a sign of low
quality, not high quality, and the best memory is made from chips that
can be easily identified as being from Samsung, Micron, Nynix, Elpida,
etc. and not overclocked.

 

[Jon Danniken]

I neglected to check the name of the Antec liquid cooler It is Antec
Kuhler H20 620

I've always wanted a water cooling setup, originally for the o/c, but
now I would like one more for the lack of noise. A few months back I
got a bigger HDD, and while this box was in my shop getting tinkered
with, I had my TV box in here to use.

The TV box, of course, was built with attention to keeping it quiet, so
I really noticed the decrease in volume while it was serving desktop duty.

Jon