3rd Party 3D IHV

[bsdgeek]

Who do you think will be most likely to successfully challenge ATi and Nvidia for the performance crown and mindshare?

[GNEP]

Where are teh BitBoys?

[SpiralDragon]

Only in my dreams... but from where i am standing there is not mutch differance ATM... GO MATROX

[Ribbit]

Right now I think S3 has the best chance, but they're going to need a lot of time. It won't be the "current" DeltaChrome chips, probably not the ones after that, but maybe the ones after that will have a shot. Or maybe they'll try really hard and get lucky sometime, like ATi with the 9700.

3DLabs I'd rate second, but they've probably sold about 6 WildcatVP cards total and so don't have the R&D money to make it into a Radeon-killer.

Matrox are still figuring out what do with themselves.

Intel are concentrating on lower-end integrated chipsets, and aren't really interested in ultimate performance right now.

PowerVR is dead. XGI have put themselves right back in the laughing-stock spot that Trident were trying to pull themselves out of.

Nappe1 is going to come on here and rip everything I've just said to shreds.

[bsdgeek]

Yes, I agree. I think S3 has the best chance of success right now, and their strategy of aiming low first will give them a real shot at grabbing marketshare. (Though had Parhelia been something truly special, Matrox's aim-High filter down might have worked.) Delta & Gamma Chrome will probably give them a solid backing to go performance with Destination Films next year, but still waiting for Deltachrome to hit the market...

3DLabs R&D money depends on whether or not it has a sugar daddy in Creative.

Intel would certainly have the marketshare and resources to go for the add-on sector, would be interesting at any rate if they tried again.

Everyone would love to see PowerVR deliver a new core, and apparently they've licensed out Series 5 to SEGA (for arcade machines and not Dreamcast2 ).

[bsdgeek]

Looks like S3 does indeed want to go for the performance crown with Destination Films. GammaChrome will likely be another midrange part like DeltaChrome.



edit: My roadmap has disappeared, I should have saved the image.

[UtwigMU]

I think Matrox has the best chance, SiS and VIA might succed with integrated and laptop chipsets, why:

You can no longer pull 3dfx or nVidia on industry (small upstart enthusiast company developing revolutionary 3d chipset and taking significant part of marketshare):
- The graphic chip design is much more demanding, you need bigger teams, better tools, you have to inovate
- The Ati and nVidia are unlike 3dfx deeply entrenched in industry with OEMs, card makers, game developers, hardware sites
- They can make cheap lowend parts that have much better driver support and compatibility

You can no longer capture lowend:
- Unlike Ati, who is still selling Rages and which had big OEM market which gave them resources to come up with Radeon, most of 2D/lowend graphics are integrated. The only time, people buy graphic cards now, is when they want to play games, they want worksation card or need superp 2D or multimonitor solution.

The days of Rages and Savages in BX chipset OEM boxes are gone.

The low end market is no longer basic 2D cards, but cards that can play lattest games at minimum settings and nVidia and Ati will not let this segment go. Furthermore this segment is much smaller than it used to be.

The S3 adn XGI parts will make it in the bottom of the bin cards, integrated VIA and SiS chipsets and lowend laptops with mentioned chipsets. In order for them to become competitive, SiS and VIA would have to burn money for a few years.

Problem is: It doesn't cost nVidia or Ati more to hire engineers in any job market (they too can open labs in Asia) and UMC or TSMC charge the same wether they make nVidia or XGI gpus.

Furthermore most of the cards are made in Taiwan or other Asian countries and again it's not more expensive to make pcbs for Ati or nVidia cards compared to XGI or S3.

The board makers are also not happy to jump on 3d party bandwagon as:
- Ati or nVidia would exert pressure
- going with less compatible, reputable and less certain upgrade cycle chipset could tarnish reputation, raise support or RMA costs (for instance I recall Hercules replacing Kyro II cards, because they didn't work in a few DX7 games although it said so in the box)
- for getting cheap lowend parts, you risk reputation and loosing ability to sell high margin high end parts

Look how manny makers make XGI or S3 chip based boards.


Would you rather have Radeon 9000 / GF4MX or would you get some S3 or XGI card of the same performance for slightly less?

Would you as system integrator rather sell systems with Ati or nVidia cards or risk higher support costs and less satisfied customers, by getting slightly cheaper S3 or XGI card?



Matrox is still the only company that has any edge in at least some area over the two and if they can expand themselves in 3D workstation market (they will need solid 3d parts for this) and 2D market, they will be able to survive. The only problem is convincing decision makers, who read Tom's that P650 is a better card than Radeon 9600.

They will need somewhat more mass market competitive product.

I also don't see 3D labs going away soon.

[bsdgeek]

I wouldn't <s>risk</s> choose higher support costs and less satisfied customers by getting a largely more expensive Matrox card.

BTW, does anyone know if new S3 got the tattered remains of old Diamond & #9 or did that stay with SONICBlue? (Read: capability to build own boards, integrate Beatles references, excellent 2D quality. Okay that was more #9 than Diamond.)

[Chrono_Wanderer]

S3 Graphics has the best chance. Recent reviews show that deltacrhome S1 has close performance to R9600 pro and FX5600... but considering the fact that NV4X is launching, this situation may change.

but then I also read they are priced like the R9000s and FX5200, its gonna be interesting IMO

[TdB]

well, powervrs devrel site have some ps3.0/vs3.0 demos, that must be because they want gamedevelopers to start doing ps3.0/vs3.0 stuff for their upcomming games.
they only reason powervr could have for wanting this must be because they plan to have a ps3.0/vs3.0 capable card out sometime in the future, those dx9 demos is most likely NOT for developers of sega argade games(I don´t think they use dx9).

of course they might just write those demos for that shaderbook, but still, the fact that they spend time on those demos must be a hint of something.

there is always the possibillity that their series5 card might never be mass produced. but their deal with sega might attract some attention from someone in the pc-market. afterall their series3 card was far from being a failure, thats gotta count for something...

[bsdgeek]

Here we go...

[Kurt]

So they're playing catchup...Nothing amazing this year...

[UtwigMU]

Let's put it this way: Would you consider buying a card from those makers?

Except Matrox and 3DLabs I wouldn't.

[SpiralDragon]

same here...
and since fps is not what i need i am stiking to matrox.... just hope for better ogl performance

[Chrono_Wanderer]

actually... i'd try S3 Delta Chrome if its cheap enough (a bit more than GFFX5200U) and has sub-R9600 Pro performance.

I won't try anything Matrox, given their current "excellent" drivers and numerous problems on the first Parhelia batches.

I will try 3DLabs too if they make a consumer card