How weak is the T&L engine of the Radeon.... click to find out....

[legion88]

IBMer wrote:

<< Now why couldn't any of you guys see it this way? >>

First, it is old news for those of us who owned a Radeon for quite some time now. I didn't think much about it as I never thought it was faster than the GTS to begin with. I was thinking that the Radeon would be in the &quot;ball park&quot; of the GeForce256, NVIDIA's first attempt at a consumer-level card with hardware T&amp;L support, when I first bought the Radeon several months ago.

Second, the dislike for NVIDIA has an impact on how people view things. You are seeing the result of that hatred. They are tired of seeing threads with NVIDIA being on top, especially since they don't like NVIDIA in the first place. You provided a link to a thread showing a NVIDIA product beating out a competitor. At some point, somebody is going to crack hearing these things so often. To me, it is just a video card, a hobby, but for others, it is a religion.

For the record, I had a GTS (Hercules Prophet II 64MB) from around May to early/mid September, Radeon (64MB Retail VIVO) from September to just last week and Voodoo5 5500 (64MB) for about a week in my current machine.

Napalm wrote:

<< ...I focused on your use of synthetic benchmarks that don't tell you jack-$hit about the real world... >>

And explain to me in detail how benchmark scores using Quake III gives me an accurate representation of the type of performance I should expect in games like the upcoming Tribes II? Did those Quake II scores from long ago convinced you that the Voodoo2/3 would be faster than the TNT2 Ultra in Quake III? The point? The point is that these so-called &quot;real world&quot; benchmarks are no better than the &quot;synthetics&quot; in predicting a card's performance in another application/game. Quake III scores are good in predicting performance of various cards in Quake III. Beyond that, you are gambling. Nothing wrong with gambling, just point it out.

 

[IBMer]

The link has been fixed.

Try clicking it now.

What I have also found out that in Running the Radeon in Athlon Optimizations instead of the D3D T&amp;L Hal that I was only 300 points from my score with the T&amp;L enabled. This does not look good being that the 3DMark2000 engine doesn't even have good 3dnow! enhancements. With the V5 you can choose for the drivers to use there own 3dnow! enhancements via Geometry excelleration or use the game engines enhancements using the Software HAL. Using GA gains me about 500points in 3dMark2000. If the Radeon also had the same 3dnow! enhancements in their drivers the score could possibly be higher than using T&amp;L which doesn't sound good to me.

 

[Soccerman]

wow IBMer you're doing pretty good for yourself, with 3 video cards on hand...

I'd like to say, that there are very convincing arguments both ways..

first of all, what about MDK2 that alot of nVidiots seem to love to point out (or did when 3dfx was still up and kicking).

doesn't the Radeon score pretty well in when you enable T&amp;L? I don't remember quite well from a while back when arguments were flying between 3dfx and nVidia whether or not the ATi Radeon did well there.

also, sure, ATi has some things going for it that the GTS does not (EMBM, which has alot more support then Cube Mapping has currently, as well as DOT3 etc). don't get into another this card is better then that card fight, I'd like to INVESTIGATE WHY becuase I'm pretty sure ATi's Radeon DOES have a T&amp;L unit onboard..

 

[IBMer]

I'd like to INVESTIGATE WHY becuase I'm pretty sure ATi's Radeon DOES have a T&amp;L unit onboard..

This is exactly what I am doing. I want to know why. I could care less if my GTS is faster than my V5 or my Radeon is faster than my GTS. I have all of them. I would like to know why in some test does the Radeons T&amp;L engine tend to be less than what you would think. I don't have a GeForce DDR to compare it too though.