[jwl812]

Well I have suspected Windows XP's SP1 to be more of a problem than a fix and now they are confirmed.
I did some benchmarking with and without SP1 installed and the numbers are intriguing.
First off, my modest system - AMD Duron 1.2gh, 512mb PC133 and Geforce2 MX400 w/64mb memory.

3dmark 2000 - Prior to SP1

3dmarks - 4919
CPU performance - 413
Game 1 - low detail - 95 fps
med detail - 70.0 fps
high detail - 36.2 fps
Game 2 - low detail - 104.2 fps
med detail - 63.7 fps
high detail - 40.7 fps

Just to highlight the whole test.

3dmark 2000 - After SP1

3dmarks - 4763
CPU performance - 398
Game 1 - low detail - 87.5 fps
med detail - 66.7 fps
high detail - 34.1 fps
Game 2 - low detail - 100.9 fps
med detail - 62.0 fps
high detail - 38.2 fps

Not much of a drop in framerates but a drop nonetheless.

In concluding, I read somewhere that SP1 cranks down some system components to offer maximum reliability and compatibilty. It seems that 3d components are some of them. I am reformatting and not installing SP1. Also, I have noticed my sound card - soundblaster live - is now popping and cracking, it didn't do that before SP1. Anyone notice any similar problems after SP1?

[Jerry Atrik]

i had the same thing until i did a clean slipstreamed install

[jwl812]

What is a slipstreamed install?

[adamvjackson]

A slipstream install is where you intergrate the installation of a service pack with the OS installation. Typically, this is done one of two ways; Either creating a new CDROM, or as a RIS image on a domain controller. Also, for the record, I too, have had the best luck with slipstream installs of XP/SP1. Do a forum search for 'Ultimate Slipstream Guide' for detailed instructions.

[jwl812]

I just completed a slipstreamed install and everything works much better now. I will be doing some benchmarks on this install soon. Everything seems to be much better. Thanks for the help.

[Jerry Atrik]

sweeet

[adamvjackson]

Cool, let us know how the benchmarks come out.

[jwl812]

Ok, the slipstream install completed without a hitch. Now here are some benchmarks. A little suprising.

Just to recap the last posting before the slipstream install......

3dmark 2000 - Prior to SP1

3dmarks - 4919
CPU performance - 413
Game 1 - low detail - 95 fps
med detail - 70.0 fps
high detail - 36.2 fps
Game 2 - low detail - 104.2 fps
med detail - 63.7 fps
high detail - 40.7 fps

Just to highlight the whole test.

3dmark 2000 - After SP1

3dmarks - 4763
CPU performance - 398
Game 1 - low detail - 87.5 fps
med detail - 66.7 fps
high detail - 34.1 fps
Game 2 - low detail - 100.9 fps
med detail - 62.0 fps
high detail - 38.2 fps

AFTER SLIPSTREAM INSTALL OF SP1

3dmarks - 4909
CPU performance - 410
Game 1 - low detail - 94.1 fps
med detail - 69.2 fps
high detail - 35.9 fps
Game 2 - low detail - 102.9 fps
med detail - 62.5 fps
high detail - 39.5 fps

The biggest difference was in raw polygons being thrown onto the screen:
High Polygon count went from 9623 mtexels/sec to 9865 mtexels/sec after the slipstream install.

Its conclusive that the slipstream install of SP1 is the way to go, but the raw framerates are a little better without SP1 being installed at all. On the other hand my system overall seems to run better with SP1. The drop in a few framerates are a small price to pay for a stable system.

When DX9 is cleaned up, I will run those too if anyone is interested so we can compare DX8.1 and DX9. I only have a Geforce 2 so the feature set of DX9 won't come into play, only if DX9 cleans up the NT Kernel for gameplay further.

[philbarnes]

wow might have to give that a go !

[Giangi]

Someone of you have compared the performance using a RAID? I have a Promise RAID SX6000 controller with a RAID 5 arraid configured with 4 disks of 60Gb. each and one spare disk. and after installed SP1 i have noticed NO better performance... anybody have more info?

Thanks...