Does Radeon have T&L?

[TigerShard]

Does Radeon have T&L? I keep hearing about the lack of T&L on KyroIIs, but what about the Radeon? Thanks

-TigerShard

 

[jobberd]

yes, the Radeon has Transform & Lighting

 

[EdipisReks]

the radeon's T&L unit is stronger than the geforce2 GTS's on paper (30 million triangles per second vs. 25 million) but in reality, the radeon's is a bit weaker. it beats the savage 2000's T&L unit, though 😉

--jacob

 

[BFG10K]

the radeon's T&L unit is stronger than the geforce2 GTS's on paper

Yes, it's rather unusual. You'd think that with Hyper-Z, its advantage on paper and memory clock speed advantage, it would be well ahead of a GTS. But it actually comes out to be weaker in the real world.

 

[MustangSVT]

weaker meaning slower right?

 

[Teasy]

<<<Yes, it's rather unusual. You'd think that with Hyper-Z, its advantage on paper and memory clock speed advantage, it would be well ahead of a GTS. But it actually comes out to be weaker in the real world.>>>

Hyper-Z doesn't have anything to do with HW T&amp;L.

<<<Does Radeon have T&amp;L? I keep hearing about the lack of T&amp;L on KyroIIs, but what about the Radeon? Thanks>>>

Every Radeon appart from the Radeon VE has HW T&amp;L.

 

[duragezic]

BFG10K wasn't talking about the T&amp;L specifically. He meant hyper-z, better t&amp;l, higher clock, etc... as a whole.

But I do agree, the Radeon *should* be much better yet the GF2's brute force approach worked better. Bwah charisma engine and all that. :|

 

[Teasy]

<<<BFG10K wasn't talking about the T&amp;L specifically. He meant hyper-z, better t&amp;l, higher clock, etc... as a whole.

But I do agree, the Radeon *should* be much better yet the GF2's brute force approach worked better. Bwah charisma engine and all that.>>>

Why should it be faster then the Geforce 2?, AFAIK it doesn't have a higher clock speed, Geforce 2 GTS is 200mhz core and memory and Radeon is 183mhz core and memory. Then there's the fact that Radeon has half the amount of pixel pipes (2 compared to the GTS's 4), it might have 3 TU's but in todays mostly dual textured games one of its TU's is usually lying dormant anyway so basically it is less then half the speed of a GTS in raw power (unless its allowed to use all 3 TU's), its only the fact that the GTS is totally bandwidth limited that Radeon can compete with it at all.

 

[TravisBickle]

ATi have long positioned there cards in the market with features, or should I say features on the sides of the box whilst not necessarily being usable in their drivers...
the third texture unit promised cute eyecandy and they must have known even the two pipelines could saturate the memory bandwidth from the start.
game designers have not fallen over themselves to support these three units, so the only advantage in it is more filtering at no cost.
most people are playing without hierarchical Z without knowing it, because it was so buggy it's disabled.
what happened to keyframe interpolation etc? are these a year down the line when the Radeon might be too slow to play such games anyway? isn't the myth of feature longevity wonderful! ask G400 owners about their bump-mapping.

 

[EdipisReks]

if hyper-z is so buggy, then why do radeon le's run so much slower than a radeon 32MB DDR card when they are overclocked to the same speed, but heirarchical z is not enabled? i'm not being sarcastic, i really want to know. as far as i know, the regular radeon's (64 MB VIVO, 32 MB DDR, AIW, and 32 MB SDR) all have hyper z enabled by default. maybe i'm wrong. speaking of clock speed, does the new SE compete well with the GTS?

--jacob

 

[MrMojo]

Actually the Radeon is very impressive, it has 2 pipelines x 3 texture units each compared to the GTS which has 4 pipelines x 2 texture units each. So the Radeon is obviously a more efficent design than the GTS since it scores so close (64mb Radeon actually outscores 32mb GTS).

 

[jobberd]

isnt it funny how a simple yes or no question can churn out 9 responses? 🙂

 

[BFG10K]

Teasy:

Hyper-Z doesn't have anything to do with HW T&amp;L.

Not directly, but it frees up memory bandwidth which T&amp;L requires.

Why should it be faster then the Geforce 2?, AFAIK it doesn't have a higher clock speed, Geforce 2 GTS is 200mhz core and memory and Radeon is 183mhz core and memory

Err, core clock speed is largely irrelevant in real world performance. It's the memory bandwidth that matters. The fact is that the Radeon with 183 MHz RAM and Hyper-Z has more memory bandwidth than a GTS with 166 MHz memory.

 

[Patgod]

<< So the Radeon is obviously a more efficent design than the GTS since it scores so close (64mb Radeon actually outscores 32mb GTS). >>



wow impressive. that's like saying my geforce 3 beats a gf2 pro 32mb.

*tumbleweeds roll by*

no one is impressed...

 

[jobberd]

well, he said that because even though it has less fillrate power, it can still outpower a Geforce2 GTS. However, A Radeon 64mb doesn't outperform a Geforce2 GTS

 

[rbV5]

EdipisReks, Hyper-z and hierarchical Z, are two different things.

 

[EdipisReks]

yeah, i realized that after i posted, but i forgot to correct my post. oh, well.

--jacob