Interview with Danny Lee from XGI

[Fox5]

Their Business model will *Never* Work.

It was working quite well for PowerVR/STMicro in the Kyro days.

Those 'days' lasted for like 6 months, and when all was said and done the kyro and powervr were not strong market brands, they didn't even exist in the market afterwards. They only did somewhat well because they offered mid range performance for low end price, XGI so far does not appear to be offering that, they haven't even seemed competitively priced so far.

XGI has the technology to bring the Graphics Memory down to 0. However, the performance won?t be competitive to our competitors.

Hasn't it been possible to go without any graphics memory with any video card, just PCI Express has the first bus that's fast enough to even consider it? Anyhow, I think they could cut out graphics memory and still compete just fine with S3.

Taipei, Taiwan, April 11th, 2005: XGI Technology, Inc. today releases its source code for Linux. To support the large and growing Linux community, XGI Technology released the source code to X.Org, XFree86.Org, and Kernal.Org for its desktop, notebook, and server products.

"The Linux community is growing strong in today's technology market. To incorporate the newest Linux development to coinciding with XGI's graphics innovations, we have decided to open our source code for Linux," said Mr. Derek LinXGI's software VP. "Our goal is to support all Linux developers so that our graphics processors can run on nearly all Linux platforms; X.Org, XFree86.Org, and Kernal.Org are the vehicles to achieve this goal."


XGI supports X.Org 6.8.2nd XFree86 4.4.0. The open source supports package installation for IA32, AMD64 and EM64T, and Frame buffer device (FBDev) supportfor kernel version 2.4.X.

Well, I suppose XGI could go for as many niche markets as possible, if you're the only or best choice in a limited market it at least gets you some name recognition.
They could also massively undercut the prices of their competitors like amd used to always do to Intel, come on xgi, release a 4 or 8 chip board and sell it for less than the cost of nvidia and ati's top end.

 

[Fodder]

Those 'days' lasted for like 6 months, and when all was said and done the kyro and powervr were not strong market brands, they didn't even exist in the market afterwards. They only did somewhat well because they offered mid range performance for low end price, XGI so far does not appear to be offering that, they haven't even seemed competitively priced so far.

If PowerVR had have had the financial and manufacturing clout of SIS and UMC behind them, perhaps they could have kept it up. XGI certainly have an opportunity here, but so far they seem to be squandering it.

 

[Hellbinder]

Those 'days' lasted for like 6 months, and when all was said and done the kyro and powervr were not strong market brands, they didn't even exist in the market afterwards. They only did somewhat well because they offered mid range performance for low end price, XGI so far does not appear to be offering that, they haven't even seemed competitively priced so far.

If PowerVR had have had the financial and manufacturing clout of SIS and UMC behind them, perhaps they could have kept it up. XGI certainly have an opportunity here, but so far they seem to be squandering it.

No...

The fact is *IF* kyro had sold well then it would not have been canned by STmicro. No one cuts off the arm that feeds it. They had plenty of manufacturing and Financial cout with STmicro. The problem is they didn't sell.

You *HAVE* to make a big splash at the high end to create the buzz that drives your product. Then you also have to have decent product. XGI has neither. Their product sucks and their Business model is doomed to failure.

 

[Fodder]

Had PowerVR had more backing, they could have brought out another generation, perhaps been more successful, and ramped up the business. Anyone expecting a newcomer to enter with a bang and be an immediate success is a fool, look how long it took NVIDIA and especially ATI to get their asses into gear. S3 and XGI's only hope of success is to work on easier midrange offerings until they get the experience and expertise to make a play for the high end. If they shoot for the high end off the bat, we end up with nosedives like the Volario Duo.

It's a similar situation with MS and the Xbox, any new entry will likely have to lose quite a bit of money to establish themselves in a market that already has such dominant players. STMicro weren't prepared to make that investment.

 

[Fox5]

Had PowerVR had more backing, they could have brought out another generation, perhaps been more successful, and ramped up the business. Anyone expecting a newcomer to enter with a bang and be an immediate success is a fool, look how long it took NVIDIA and especially ATI to get their asses into gear. S3 and XGI's only hope of success is to work on easier midrange offerings until they get the experience and expertise to make a play for the high end. If they shoot for the high end off the bat, we end up with nosedives like the Volario Duo.

It's a similar situation with MS and the Xbox, any new entry will likely have to lose quite a bit of money to establish themselves in a market that already has such dominant players. STMicro weren't prepared to make that investment.

Better planning could have helped the xbox, and many other business ventures.
And maybe not attacking the high end could work....if they play to a different market entirely. Offer something ATI and Nvidia deem too insignificant to offer and it might take off.

BTW, S3 Unichrome is pretty bad too, yet it doesn't get nearly as much flak. Maybe because it's affordable and generally integrated graphics or a laptop chip?

 

[Fodder]

Unichrome is absolute bottom of the heap and S3 make no allusions to the contrary, but Deltachrome and Volari were actually supposed to be good. :?

 

[Mendel]

Had PowerVR had more backing, they could have brought out another generation.

AFAIK, Kyro WAS the another generation. Or third generation maybe? I have good memory of the first PCX parts which arrived roughly at the same time with Voodoo graphics, but failed to deliver bilinear filtering, which turned out to be a feature not acceptable to miss. In next generation they had something... well I just recall it finally had bilinear but failed to have any good translucency/alpha features.

Then along came the Kyro. Again behind in features, missing the all important and rediculously hyped hardwre T&L support, and a bit too low clocks, Kyro 2 was clocked a bit higher, but apparently too little too late. Kyro 2 SE apparently had some sort of T&L emulation, but I'm not too familiar with that tech.

 

[Fox5]

Had PowerVR had more backing, they could have brought out another generation.

AFAIK, Kyro WAS the another generation. Or third generation maybe? I have good memory of the first PCX parts which arrived roughly at the same time with Voodoo graphics, but failed to deliver bilinear filtering, which turned out to be a feature not acceptable to miss. In next generation they had something... well I just recall it finally had bilinear but failed to have any good translucency/alpha features.

Then along came the Kyro. Again behind in features, missing the all important and rediculously hyped hardwre T&L support, and a bit too low clocks, Kyro 2 was clocked a bit higher, but apparently too little too late. Kyro 2 SE apparently had some sort of T&L emulation, but I'm not too familiar with that tech.

Well, there was the first gen that didn't have bilinear filtering.(or so one article says, I've never found anything else to support that it didn't) It didn't have glide either, which was the best api at the time.
Neon 250 was released too late to be competitive and had buggy drivers, particularly in opengl. Performance in games like quake was really good, performance in games like freespace was horrible.
Then I think Kyro and Kyro 2 were released at the same time, but kyro was a low end part and kyro 2 only a mid range part. I think T&L emulation was available on all the kyro parts, but may have been lacking in performance and compatibility.(3dfx's cards had it too, but it only worked with dx7, dx8 completely broke compatibility and 3dfx was gone by then) I don't think the Kyro 2 SE ever came out, but it was just to be a higher clocked Kyro 2.(and pretty much the reason I avoided the kyro 2, kyro 2 se was supposed to be out shortly after and not cost too much more)

 

[Fodder]

I vaguely remember after Tim Sweeney had his rant about TBR's being unable to support T&L in hardware, there were some murmurs out of PowerVR about the next series being perfectly capable (and more).

 

[Fox5]

I vaguely remember after Tim Sweeney had his rant about TBR's being unable to support T&L in hardware, there were some murmurs out of PowerVR about the next series being perfectly capable (and more).

Why wouldn't they be able to? Why does it matter if T&L is done by the graphics chip or by the cpu?

 

[Fodder]

Why wouldn't they be able to?

You'd better ask Sweeney.

 

[Fox5]

Why wouldn't they be able to?

You'd better ask Sweeney.

But what about Elan?