New Roadmap of S3 Graphics

S

Stefan

Newcomer
http://www.chrome-center.com/index.php?option=com_content&task=view&id=78&Itemid=63

GammaChrome S19 and S14 are canceled, but with Chrome S2X (S27, S25 and S23) a SLI like technology called Multi-G is coming. 8x1 and 700 MHz in November/December.

And some decent Notebook-Chips are also set to arrive in late 2005/early 2006.

Destination Films is scheduled to arrive in Q3 and Q4 2006!

artroad.gif
 
2 quads @ 700MHz could do fine under presuppositions by the end of this year.

Destination Films will be a different story since it'll be a WGF2.0 GPU.
 
I think it's quite fascinating to see S3 still doing something (or at least trying to sell the image of), when in fact, from it's recent track record, the whole VIA conglomerate seems rather defunct when it comes to R&D. Good for them! And 5.6 MTex/sec don't sound bad for a mainstream card either, provided they have a decent shader core, of course.

BTW: Is S2x SM2 or SM3?
 
What's the advantages of any of the S3 line? Seems they are always behind and not up to par as far as performance and drivers. I havent really had much experience with them though.
 
This is perhaps somewhat off-topic, but do any of you know if S3 is making sales and money? I don't see their products around much.
 
Probably not, but thn the benefits are in other areas. With ATI and NVIDIA increasingly encroaching on VIA's space in the chipset market, they will need an integrated solution to keep them competetive.
 
S3 chips have some positive aspects:

- correct trilinear and anisotropic filtering
- no other driver cheats
- they consume only low power in comparison
- HDTV capable

And of course also some negative points:
- Only SSAA, limited to 2x [5x and 9x are possible but only at 640 x 480 afair)
- DeltaChrome was inefficient, GammaChrome is better in that way, but still not optimal
- bad market penetration
- the drivers have some weak points, but the become better and better with each release

BTW: Chrome S2X is SM 2.0 only based on the informations known until now. SM 3.0 ist quite unlikely because Chrome S2X is still baes on the Columbia design and the changes would be too big for that. The shader power of GammaChrome is like ATI, sometimes better, sometimer worser. Chrome S2X is much likely the same.
 
Last edited by a moderator:
correct trilinear and anisotropic filtering
Click to expand...

How many texels does it fetch for fast trilinear?

***edit: I hope they go for Multisampling this time.
 
Stefan said:
S3 chips have some positive aspects:

- correct trilinear and anisotropic filtering
- no other driver cheats
- they consume only low power in comparison
- HDTV capable

And of course also some negative points:
- Only SSAA, limited to 2x [5x and 9x are possible but only at 640 x 480 afair)
- DeltaChrome was inefficient, GammaChrome is better in that way, but still not optimal
- bad market penetration
- the drivers have some weak points, but the become better and better with each release

BTW: Chrome S2X is SM 2.0 only based on the informations known until now. SM 3.0 ist quite unlikely because Chrome S2X is still baes on the Columbia design and the changes would be too big for that. The shader power of GammaChrome is like ATI, sometimes better, sometimer worser. Chrome S2X is much likely the same.
Click to expand...
Yeah, but aren't they sort of an unknown commodity? :???:
 
digitalwanderer said:
Yeah, but aren't they sort of an unknown commodity? :???:
Click to expand...

That doesn't mean though that they can't work themselves up in the future for higher market penetration when it comes to notebooks and IGPs f.e. With VIA's backup it shouldn't be too hard it just takes persistance and repeated (increasing) successes.
 
Stefan said:
S3 chips have some positive aspects:

- correct trilinear and anisotropic filtering
- no other driver cheats
Click to expand...
Is this right? AFAIK, ATI's bilinear AF and trylinear followed from some employees' emigrations from S3, and Xbit or iXBT exposed some driver-directed IQ shortcuts taken in some popular benchmarks. (In other words, no one's immune.) Maybe I'm confusing S3 with XGI as it relates to the app detection.

Anyway, I'm excited by their continued development. Who knows, serendipitous things may spring from it and positively affect the entire 3D market.
 
Last edited by a moderator:
Ailuros said:
How many texels does it fetch for fast trilinear?

***edit: I hope they go for Multisampling this time.
Click to expand...
If you've read my article, you'd know that fast tri fetches 16 texels.
Multisampling will be implemented in destinantion films. Chrome S2x is just a 2 Quad Gamma Chrome. Its further name was S19. Just another lable for the same chip.
The increase of arithmetical efficiency was not to bad from delta chrome to gamma chrome (got an article about that in my pipe). But per clock and pipe its way behind ATI's R3xx efficiency.
 
Dave Baumann said:
So, you have an inate understanding of the architecture of those pipes already?
Click to expand...


No, but I highly doubt that their single pipe will be as fast as two nV or ATI ones. X800 or the [edit]12-pipe 6800[/edit] will be the "lower-end" by the end of the year I'd say.
 
Last edited by a moderator:
Pete said:
Is this right? AFAIK, ATI's bilinear AF and trylinear followed from some employees' emigrations from S3, and Xbit or iXBT exposed some driver-directed IQ shortcuts taken in some popular benchmarks. (In other words, no one's immune.) Maybe I'm confusing S3 with XGI as it relates to the app detection.

Anyway, I'm excited by their continued development. Who knows, serendipitous things may spring from it and positively affect the entire 3D market.
Click to expand...
I haven't any knowledges that there is cheating at in benchmarks! And I've read all reviews of DeltaChrome and GammaChrome. There is even an official announcement from S3 Graphics that they aren't using any "optimizations" in benchmark programs!
I really think you are confusing with XGI: They use optimizations/cheats for nearly every program/game/benchmark with reduced image quality.

S3 has a technology called fast trilinear which confuses the theoretical programs for AF and therefore the shwon results are wrong sometimes, but it's working right and produces full IQ.
 
Stefan said:
S3 has a technology called fast trilinear which confuses the theoretical programs for AF and therefore the shwon results are wrong sometimes, but it's working right and produces full IQ.
Click to expand...

No it does not. The AF is broken which makes it completely unusable, the IQ is inferior to everything i have seen so far.
 
You must log in or register to reply here.
Back
Top