GeForce FX Go5600 - 3D Mark Scores

[Guest Vandaahl]

no man its the right link i got 1495 with the drivers supplied by acer (they only have 1 version :) )

and the +1000 points score is from pieter's drivers

[MicroChip]

This is quite normal for 3DMark 2k1 to be slower with 5x.xx than 4x.xx.

5x.xx drivers concentrate on DX9, so you'll find that 3DMark 2k3 improves with 5x.xx over 4x.xx.

So for max speed use 4x.xx drivers for DX8.1 or lower and 5x.xx for DX9 and higher.

Hope this helps,

Pieter.

This is good to know, I've got the ZD7000 with the 128mb FX go5600, 1 gig of ram as well, and my 3dmark 2001se dropped significantly from the stock drivers to some of the newer modded ones, yet when I ran some of the UT2004 benchmarks as posted somewhere else here, some were faster with older drivers, some with newer. I've got 3dmark 2003 around, I'll see what's fastest.

Right now I'm running 54.01, which I think came from Windows Update and getting a hair under 10,000 3dmarks. The previous drivers which came on the laptop was just a hair over 10,000. No tweaking, no overclocking.

I still don't know what the best driver to use is, and the biggest problem I have is every time I put a new driver on, I run rivatuner just to see what I've got, and sometimes like the NVIEW version won't match the driver version, pita.

What's the best way to strip off the drivers, reboot, and put a new one on, as clean as possible? Also, every once in a while I strip off the driver and end up with VGASAVE as a device installed, and can't get it to install an NVIDIA driver...

Maybe we should start an "Official ZD7000" thread?

MC

[mobilenvidia]

MC,

Have you tried 44.03 if you want to break 10K this will do it.

It's the fastest DX8.1 driver (not necesarily the best)

Post your results and I'll post it up on my website's performance page (top 3)

I'm only interested in NON overclocked scores 3DMark 2k1 and 2k3.

This applies to any body with an nVidia GPU no matter what, I'd like to get a database on all GPU's.

Uninstalling use Control Panel, 'Uninstall programs' then the video driver.

My unistall section in the INF is quite comprehensive now.

Enjoy,

Pieter.

[MicroChip]

Cool thanks, I'll do some benchmarking with assorted driver versions this weekend. I'm less interested in boosting my 3dmark2001 score and more interested in my 3dmark2004 score, and how the machine plays UT2004 and Battlefield Vietnam, eagerly awaiting Doom III.

Here's a question though, when I had the stock drivers from HP, and the ones that Windows Update added, Rivatuner reported the core and memory clocks at one set of speeds (I'm at work, wrote them down at home), and when I installed 56.64 with your 20.22 inf, with the settings for overclocking off, I think the defaults were lower, is that possible? So when you say you only want non-overclocked scores, how will I know? With the windows update driver (was it like 54.something?) my 3dmark03 was a hair over 2500. With 56.64 and no overclocking it dropped, but I could ramp up the clocks not too hard and go to almost 2700.

So with the whole non-overclocked only scores, I'm just not sure is every version of the driver using the same default clocks?

Did I confuse anybody or sound like an idiot? I'm ok with that

MC

Also, I meant 3dmark2003, not 2004 (I don't think there is a 2004 yet).

Lastly, as an interesting point, there's a GREAT cpu benchmarking article at Techreport which does a great job comparing every CPU you'd want to know about now. Here's the thing, they test all the p-4's, using 1 gig of ram and an SATA drive, and here's the key, a Radeon 9800 pro with 256mb ram.

With that testbed, the same CPU and same amount of ram as I have, the differences being the Go5600 128mb and a 4200 rpm 80 gig drive vs the Radeon 9800 256mb and SATA drive, they got over 5700 3dmarks using 2003.

Is a 256mb Radeon 9800 that much faster than a 128mb Geforce 5600? 5700 3dmarks vs 2700? Wow, that's a huge difference.

I just found that interesting.

MC

Edited May 3, 2004 by dontknow