Information

Dear forum members,
We are delighted to inform you that our forums are back online. Every single topic and post are now on their places and everything works just as before, only better. Welcome back!



Discussion

Discussion on Article:
Counter Strike: Source: Performance Preview

Started by: dan2097 | Date 08/21/04 02:49:33 AM
Comments: 47 | Last Comment:  11/26/07 10:43:59 AM

Expand all threads | Collapse all threads

[1-6]

1. 
Would you be able to elaborate a bit more on your testing methodology?

Specifically did you use control panel af or application af? Also I expect you used trilinear optimizations on the Nvidia cards right? Did you also use anisotropic filtering optimizations?

Im doing a bit of investigation to try and find out why sites are coming out with such different results

Dan2097 (Futuremark forum)
0 0 [Posted by:  | Date: 08/21/04 02:49:33 AM]
Reply

2. 
Anton thanks a lot for the quick reply. Im havnt really looked at your custom benchmarked cs source benchmarks as as you say you can directly compare to other sites. I think the status of anistropic filtering optimizations could be the key to the differences between reviews as many sites probably use the default setting of anisotropic filtering optimizations off.

From a chat with someone in a 6800 when anisotropic filtering optimizations are enabled you get bilinear on texture stages 1-7 regardless of whether af is forced through the control panel or done by application preference. This should be reproducable under d3d af tester. ATIs on the other hand is undisablable but only acts when af is forced from the control panel.

My current opinion on this is possibly Nvidias anisotropic optimizations should only actually be enabled when doing comparisons using control panel af.

Does the X800 pro lead the 6800GT or do they ~tie when the 6800GT does no have anisotropic optimizations? However as you say and I tend to agree there is no difference between the x800s and 6800s image quality with anisoptopic optimiaztions on, thus even if I am correct there is a strong case for your comparison despite the possibly unequal workloads.

Also a bit unrelated, but do the gffxs actually get FSAA in cs source/the stress test. Techreport claimed they didnt :-S
0 0 [Posted by:  | Date: 08/21/04 05:46:41 AM]
Reply
- collapse thread

 
Try changing the name of the .exe file to see if we have application spefic optimizations yet.
0 0 [Posted by:  | Date: 08/21/04 06:31:35 AM]
Reply

3. 
nice test, but what about the users NOT owning such an awesome system? its clear that css and hl2 will run fine on a system like amd64 3400+, 6800 or x800 card, 1GB LL-RAM...

what about systems like amd 2500+, geforce4ti or radeon 9500, 512MB RAM???
0 0 [Posted by:  | Date: 08/21/04 05:55:34 AM]
Reply

4. 
Is the 5950 using the full DX9 codepath or a mixed mode? and could you do some benchmarks of lower end hardware? :
Gefoce 4 TI
Radeon 8500/9200/9000
9600/9500
Athlon XP 2500/3200
Pentium 4 2.4ghz/2.8ghz
512 RAM
0 0 [Posted by:  | Date: 08/21/04 06:29:25 AM]
Reply

5. 
i noticed on some of the test's you had af set 8/16
which card's were running 8x af and which were running 16x?
0 0 [Posted by:  | Date: 08/21/04 05:09:03 PM]
Reply

6. 
I thought NVIDIA optimizations where already gone?
If they don’t (or you didn’t disable them) why you didn’t you enabled performance mode on Ati 9x00 that would give much the same image quality of NVIDIA, but even that is not enough....

It’s not CORRECT that you put all the cards in the same chart if there are different render paths. You should have forced somehow all to render DX9 or used different color in charts or spare them.

0 0 [Posted by:  | Date: 08/23/04 03:15:39 AM]
Reply

[1-6]

Back to the Article

Add your Comment