Printable View
Just curious, but why don't you guys use UT in your myriad video card reviews and guides? With the amount of textures required in the game, wouldn't it make a good option to show the more memory-intensive-game side of things? Again, just curious.
------------------
Bollenator
The gods do not protect fools; fools are protected by more capable fools. -Larry Niven
Hi Bollenator,
We get this question a lot and whilst we'd truly like to be able to use it (great game that it is), it's proved on many occasions to be a very poor benchmark. It'll show huge disprepencies when compared to other benchmarks. Also it's a pretty flat one at that and mostly dependant on teh number of bots running around at one time. A Savage 2000 can in most cases turn out the same fps as a GeForce2... Not realistic https://www.sharkyforums.com/images/.../2005/06/6.gif
Quote:
Originally posted by bollenator:
Just curious, but why don't you guys use UT in your myriad video card reviews and guides? With the amount of textures required in the game, wouldn't it make a good option to show the more memory-intensive-game side of things? Again, just curious.
------------------
Cheers
Editor-in-Chief & Founder
-------------------------
www.sharkyextreme.com
www.sharkygames.com
----------------------
Thanks for flying with us, we know you have a choice...
Thanks for the clarification, alexross. Question answered. https://www.sharkyforums.com/images/.../2005/06/5.gif
------------------
Bollenator
The gods do not protect fools; fools are protected by more capable fools. -Larry Niven
No worries, glad that settles it https://www.sharkyforums.com/images/.../2005/06/5.gifQuote:
Originally posted by bollenator:
Thanks for the clarification, alexross. Question answered. https://www.sharkyforums.com/images/.../2005/06/5.gif
------------------
Cheers
Editor-in-Chief & Founder
-------------------------
www.sharkyextreme.com
www.sharkygames.com
----------------------
Thanks for flying with us, we know you have a choice...
Hmmm: are Sharky and alexross actually the same user https://www.sharkyforums.com/images/...2002/02/11.gif?
Stoo
Stoo yes they are, I mean yes I am https://www.sharkyforums.com/images/.../2005/06/6.gif
Quote:
Originally posted by stoo:
Hmmm: are Sharky and alexross actually the same user https://www.sharkyforums.com/images/...2002/02/11.gif?
Stoo
------------------
Cheers
Editor-in-Chief & Founder
-------------------------
www.sharkyextreme.com
www.sharkygames.com
----------------------
Thanks for flying with us, we know you have a choice...
Sharky,
I completely understand why the specific tools (3dmark, q3, ect) are used for bench marks. However UT is a very popular game as is CS and some others. It would be nice to see how certain cards do on more than Q3, MDK2 and Expandable. I know that your trying to bench mark the card and CPU limited games are not good for that. But how would you feel as a consumer if you only wanted to play UT and you spent the extra $$$ to get a GF2 (again I know that there is no contest between these two video cards when it comes to speed but there is a point here) only to find out that you did not need to spend the $$$? It would be nice to see how these cards perform in other games. However you folks can not test every game out there.
I personally feel that the way benchmarks are done know is flawed. They rely to highly on fill rate limited tests (ie Q3) only. No way to measure visual quality (ok that is very subjective). I feel that some CPU limited games should be used as well. Why? Because the Video card is only part of the system that one need in order to use it. The game it is running also has as much if not more effect on the outcome as the card dose. I think almost as much info that can be gained from isolating it in fill rate limited tests, also can be gained from seeing how it works with CPU limited cases. If every game out there was fill rate limited, the today's standards are great. But they are not. Two prime examples are CS and UT. If you look at the GF2 vrs V5 cards and that debate you will see that in CS, the scores are very close with a slight edge maybe going to the GF2, and UT the V5 wins. Heck the V5 even beats the GFU Ultra in UT!!! Now looking at all of the "common" benchmarks the GF2 kills the V5. So by today's common benchmarks that show us that the GF2 is greater than the V5 and that is the card we should have. Wrong. In CPU limited games there is little difference and now you have to base the cards on other things (FSAA, NSR, T&L EBM, price, stability, ect...) Looking at only fill rate numbers of is only looking at part or the picture.
I just showed 2 very popular games (which are the #1 and the #3 played on-line) where there is no difference. Yet by todays benchmarks there is clrealy a card that is faster. To me that is a major flaw in the system. If it were two games that we never heard about, ok it is one of those one-in-a-million things (well that would be two-in-a-million). But now that I am looking into flight/space sims and sports games I am finding that they mostly CPU limited = similar performance. It seems that its the 3d shooters that are the big fill rate games.
Also 3dmark is way too biased for T&L for today's games. 3dmark well tell you a lot about your video card, but it relates very little to real world gaming performance. Again look at the CT/UT examples. The GF2 cleans up on the V5 in 3dmark, yet it has the same or lower performance in D3D vrs the V5 in those games. Dose not make sense as 3dmark showed one to be faster and yet in these 2 case there is little if any difference?? I don't trust 3dmark after I found out it bases is scores only on the T&L tests, other tests like the fill rate tests do not factor into the final score :-(
And trying to predict what the furture will hold is practially impossible. We can see what is comming in general to a certian point (for example we all know T&L is comming) but to what degree? how much? and what the drawback will be? Nobody knows.
Do you think todays benchmarks are good? fair to both CPU and fill rate limted results? Predecit the furture (like 3dmark was suppose to)?
I sure hope the reviewer sites can be opened minded about this and start included the fact that in CPU limited games you have similar performance in all of the cards, or benchmarking some of the popular CPU limited games to show this. I have seeen way too many reviews where they say one is the hands down winner however in CPU limted games, there is no hands down winner.
You folks have went out of your way to comment on FSAA and other things that these cards have to offer and how that can be a big thing in a buying decision. And I commend you for that. But more work needs to be done.
[This message has been edited by joeb (edited September 15, 2000).]