Overclock doesn't work

[dorkus.erectus]

I installed an Arctic-Cooling VGA Silencer today with the intention of
OC'ing my Sapphire 9800Pro/128MB card.
After several hours of testing, first with Powerstrip then ATItool (I'm
using 3Dmark'01, Quake3, UT2003 and Aquamark as benchmarks) I get ZERO
performance gain.
My stock settings are378core/338mem. I get the same results in ALL tests up
to 425/338 (5MHz increments between tests. No artifacts).
This makes no sense to me. Can someone tell me what I'm doing wrong?

XP2600+Barton@ 1.96GHz
512MB PC3200-DDR
MSI KT3-Ultra @ 170FSB

Rick

 

[Sham B]

Are the settings actually going in? When you change the values, if you go back into the settings,
are they still as you left them, or do they reset?

Sometimes you will find that you set an OC value in Powerstrip, but when you look at the values
again, they have reset. I have no idea what causes this, or why it happens, but I know that some
individual cards are more likely to do it than others (I have had two cards, and the last one, a
9500 pro, did it often). Normally, I just set the values again, and all is well.



S

 

[dorkus.erectus]

">

Sham B" <[email protected]> wrote in message

Are the settings
actually going in? When you change the values, if you go back into the
settings,

are they still as you left them, or do they reset?

Sometimes you will find that you set an OC value in Powerstrip, but when you look at the values
again, they have reset. I have no idea what causes this, or why it happens, but I know that some
individual cards are more likely to do it than others (I have had two cards, and the last one, a
9500 pro, did it often). Normally, I just set the values again, and all is well.

Yes the values are going in.
I learned early in my attempts that both Powerstrip and ATItool aren't as
easy as they should be but I figured them out.
Since my first post, I tried 3DMark '03 and now my puny brain is at
OVERLOAD!!
My score went from ~5300 (378/338)up to ~5700 (420/338) but all the other
benchmarks remain the same.
WTF?

Rick

 

[patrickp]

I installed an Arctic-Cooling VGA Silencer today with the intention of
OC'ing my Sapphire 9800Pro/128MB card.
After several hours of testing, first with Powerstrip then ATItool (I'm
using 3Dmark'01, Quake3, UT2003 and Aquamark as benchmarks) I get ZERO
performance gain.
My stock settings are378core/338mem. I get the same results in ALL tests up
to 425/338 (5MHz increments between tests. No artifacts).
This makes no sense to me. Can someone tell me what I'm doing wrong?

XP2600+Barton@ 1.96GHz
512MB PC3200-DDR
MSI KT3-Ultra @ 170FSB

Rick

That's as it should be. An Arctic Silencer will not give you any
performance gain at the same settings you used before.

What it may do is allow you to overclock your card to higher settings.

patrickp

(e-mail address removed) - take five to email me

 

[Navid]

I installed an Arctic-Cooling VGA Silencer today with the intention of
OC'ing my Sapphire 9800Pro/128MB card.
After several hours of testing, first with Powerstrip then ATItool (I'm
using 3Dmark'01, Quake3, UT2003 and Aquamark as benchmarks) I get ZERO
performance gain.
My stock settings are378core/338mem. I get the same results in ALL tests
up
to 425/338 (5MHz increments between tests. No artifacts).
This makes no sense to me. Can someone tell me what I'm doing wrong?

XP2600+Barton@ 1.96GHz
512MB PC3200-DDR
MSI KT3-Ultra @ 170FSB

Rick

Are you saying that you cannot overclock to a higher clock rate after VGA
silencer, or are you saying that your benchmark scores are the same as
before at the same clock rates?

If your scores have not gone up at the same clock rates, I am not surprised.
The only thing that can increase your score is increasing the clock rates.
The objective of using the silencer is to cool down the core better so that
it can be clocked at a higher rate and still be stable.

If you are saying that the card goes unstable, as you gradually increase the
clock rates, at the same rates as it did before, that is a reason for
concern!
The only thing I can suggest is to make sure that the heat sink makes good
contact with the chip. The silencer heat sink is very big. If anything is
in the way, like a capacitor on the board or the PCI bracket, it may be held
slanted in such a way to make contact with the chip only in one edge. That
will reduce heat transfer.
Also, make sure to tighten the two screws one at a time and gradually so
that the sink is pressed against the chip on all sides, and not on one side
more than the other.

 

[dorkus.erectus]

I installed an Arctic-Cooling VGA Silencer today with the intention of
OC'ing my Sapphire 9800Pro/128MB card.
After several hours of testing, first with Powerstrip then ATItool (I'm
using 3Dmark'01, Quake3, UT2003 and Aquamark as benchmarks) I get ZERO
performance gain.
My stock settings are378core/338mem. I get the same results in ALL tests up
to 425/338 (5MHz increments between tests. No artifacts).
This makes no sense to me. Can someone tell me what I'm doing wrong?

XP2600+Barton@ 1.96GHz
512MB PC3200-DDR
MSI KT3-Ultra @ 170FSB

Rick

I thought my OP was clear enough but maybe not.
Forget about the VGA Silencer for a moment.
The real problem is that when I overclock the core from the stock 378MHz to
425MHz or anywhere in between, the benchmarks remain the same.

The silencer is mounted properly.

Rick

 

[Navid]

I thought my OP was clear enough but maybe not.
Forget about the VGA Silencer for a moment.
The real problem is that when I overclock the core from the stock 378MHz
to
425MHz or anywhere in between, the benchmarks remain the same.

The silencer is mounted properly.

Rick

I can think of two things that can cause this (unless we are overseeing
something):
1- The overclocking is not really working and you are always testing at the
same rate.
2- The bottleneck (limiting the 3D performance) is not the graphics card but
the CPU. This is remote. Because, even if the CPU is the bottleneck, you
should still get a very small improvement as you increase the GPU clock
rate.

Underclock your graphics card as well. Also under/overclock your CPU and
benchmark at different configurations.

 

[Navid]

I can think of two things that can cause this (unless we are overseeing
something):
1- The overclocking is not really working and you are always testing at
the same rate.
2- The bottleneck (limiting the 3D performance) is not the graphics card
but the CPU. This is remote. Because, even if the CPU is the bottleneck,
you should still get a very small improvement as you increase the GPU
clock rate.

Underclock your graphics card as well. Also under/overclock your CPU and
benchmark at different configurations.

I just remembered that I tried RADLINKER and got confusing results. I could
not figure out what was wrong. If I purposely caused a freeze by
overclocking the GPU too much, my 3DMark03 score would go down (at the same
stable rate as before). The only way to restore my score was to use the XP
system restore. I switched completely to ATITOOL. I am happy with ATITOOL.
You may want to try different overclocking tools until you are happy with
one.