Is 48 hours of Prime 95 stable enough? - 2500+ @ 2.2GHz (1.75v)

Hi,

Just got some more spare time to play with my
*alt.comp.hardware.overclocking.amd* machine :P, yes of course all your guys
here helped me spec up this beast, with an aim to overclock!

I have been slow to whack up all the settings and see what kind of overclock
I could achieve, just that this was my first AMD build in the past 4 years
and I needed to feel happy that everything was sweet @stock settings.

Anyway to make my CPU pass prime 95 torture test when running at 2.2GHz
(11x200) I needed to increase the Vcore to 1.75v. I have been running
Prime95 for nearly two days now and the Max load I saw was 51°c (during a
very sunny day in London).

It's still burning away right now, do you guys think that 48 hours is
sufficient test that my overclock is stable? I'm tempted to leave it
running and throw in Super-Pi and maybe a radeon demo to max it, or can I
just get on with playing some games now :P

Wayne ][

Wayne Youngman wrote:
Hi,

Just got some more spare time to play with my
*alt.comp.hardware.overclocking.amd* machine :P, yes of course all
your guys here helped me spec up this beast, with an aim to overclock!

I have been slow to whack up all the settings and see what kind of
overclock I could achieve, just that this was my first AMD build in
the past 4 years and I needed to feel happy that everything was sweet
@stock settings.

Anyway to make my CPU pass prime 95 torture test when running at
2.2GHz (11x200) I needed to increase the Vcore to 1.75v. I have been
running Prime95 for nearly two days now and the Max load I saw was
51°c (during a very sunny day in London).

It's still burning away right now, do you guys think that 48 hours is
sufficient test that my overclock is stable? I'm tempted to leave it
running and throw in Super-Pi and maybe a radeon demo to max it, or
can I just get on with playing some games now :P

I'd say 48 hours without a single error should be fine Wayne. A 2500+ at
1.75V, 51°C and 2200MHz is nothing to sneeze at. Prime 95 stresses the CPU
and memory. 3DMark2001SE is a good one to run to stress the video. The 32
million digit test in Super Pi is also good for testing CPU/memory, though
it's probably not necessary at this stage.

"S.Heenan" wrote
I'd say 48 hours without a single error should be fine Wayne. A 2500+ at
1.75V, 51°C and 2200MHz is nothing to sneeze at. Prime 95 stresses the CPU
and memory. 3DMark2001SE is a good one to run to stress the video. The 32
million digit test in Super Pi is also good for testing CPU/memory, though
it's probably not necessary at this stage.

Hi,

thanks for feedback, that's good to know. Towards the end of the 48hour
test period I did indeed through in the Super PI 32 mill test, so it was
running both of them simultaneously, all good. Just to push it I ran my ATI
Gargoyle screen saver, so now all three were running. All worked fine! 0%
errors. .

Funny how 0.25v can make the difference between total rock solid (1.75v) and
fail Prime95 fast (1.725v). Nick Salmon mentioned that it may be safer to
bump it up another *notch* (1.775v) to give it even more stability, I will
bare that it mind. . .

Now if only I could get my chip to 2.3-2.4GHz like you Pro's, he-he 2.2GHz
does indeed feel like a rocket-ship, so either way it's all good

I think I may loop 3DMark2001SE or something like that to double check the
3d side of things. .

Wayne ][

Wayne Youngman wrote:
Hi,

thanks for feedback, that's good to know. Towards the end of the
48hour test period I did indeed through in the Super PI 32 mill test,
so it was running both of them simultaneously, all good. Just to
push it I ran my ATI Gargoyle screen saver, so now all three were
running. All worked fine! 0% errors. .

Funny how 0.25v can make the difference between total rock solid
(1.75v) and fail Prime95 fast (1.725v). Nick Salmon mentioned that
it may be safer to bump it up another *notch* (1.775v) to give it
even more stability, I will bare that it mind. . .

Now if only I could get my chip to 2.3-2.4GHz like you Pro's, he-he
2.2GHz does indeed feel like a rocket-ship, so either way it's all
good

I think I may loop 3DMark2001SE or something like that to double
check the 3d side of things. .

Or even 1/10th of .25V Yes, it can and often does make a difference. Due
to the thread priority of different applications, I'm not sure running P95
and Super Pi, at the same time provides the same test as running them
seperately. At any rate, try something to stress the video card,
3DMark2001SE, Quake 3 Demo, etc. Then pat yourself on the back for a job
well done.

--
Winerr 012 - Cash Underflow - Credit Card Number Will Be Assimilated