STM pulls the plug on Kyro

Why? VIA are massive in the PC motherboard market (my last two mobo's have been VIA and there both massively popular and very high quality) and surely they could be just as big in the PC graphics card market too. Why not?

They could put Kyro based chips in some of there new mobo's which would ensure a bigger TBR userbase and because there a company that's focused on the PC market they could enter the PC graphics add in board market quite aggressively if they have the right tech (which is where PowerVR 3, 4 and 5 comes in). Also they have a good relationship with TSMC and have access to there 0.13 micron process. It sounds like it could be a great move for IMGTEC IMO.

<font size=-1>[ This Message was edited by: Teasy on 2002-02-11 21:22 ]</font>

 

Because IMO, no chipset maker is really interested in the (relatively speaking) low-volume, high-performance products. They are interested in making chipsets. Low cost / high volume is the priority, and performance / features comes second. In other words, I doubt Via or Sis would buy STM and have them churn out high-level add-in board graphics chips. They'd probably have those engineers work on integrating the graphics technology with their chipsets.

I agree that someone like Via/Sis partnering for IMG Tech could be great for IMG TEC. I just don't see it as great for the gaming consumer, because I doubt they would focus on high-performance add-in boards.

 

Teasy,

VIA being a large player in the motherboard market does not stand for much. As ST has already proven, having a large company backing up the technology does not mean they will press forward. The pace in the motherboard market is incredibly slow compared to the graphics market. And slow is the last thing we want.

They also already own S3 Graphics. If they showed any interest in the graphics area, they would have tried to do something with the technology. My guess is that they didn't even have an interest in S3's graphic technology in the first place. Instead, S3's Intel license, which allows them to build Intel motherboards, is what they were after.

You make it sound so easy.

-dksuiko

 

A very good point. This had crossed my mind. If they were serious we would have seen Savage 2000's big brother a LONG time ago.

 

I wonder why VIA cancelled Columbia?

 

Source, links, details ?

 

All this speculation is getting us anywhere.

[crazy mode]
Maybe we should all group together and buy out STM graphics division!

Hmm, there are about 300 of us in this forum. Let's say STM graphics division is worth $20 million. All it would need from each of us is $66,666 (hmm, spooky number).

Then we can bring out a high end PVR chip for ourselves...
[/crazy mode]

 

We already have a fund drive going.

So far in addition to the savings I'm accumulating for my own house, the following "donations" have been made in this thread toward the cause.

* Someone's sale of a "high-end Mercedes"
* $1500 cash
* A Dollar
* Lint

 

Well in that case Joe I'll "donate" my 3D Prophet 4500 64MB card to the fund drive.

 

They can integrate PowerVR Series 3 and make mid level cards with PowerVR Series 4, they could also go for the highend with Series 5. That last part is unlikely but its likely they would do more then just integrated chips and there not going to use Kyro III as an integrated chip because Kyro II 4800 is perfect for that.

Then who in your opinion would? It seems VIA are the best option, and even if they didn't focus on the highend they could at least aggressively focus on mid-range chips which STM didn't do.

The difference being that STM were not a company focused on producing PC products, VIA are focused on the PC market.

Maybe S3's tech was just poor, I've never seen decent 3d tech from S3. Also I remember Wavey saying that its been said to him before that VIA didn't realise that S3 didn't have any future graphics projects when they bought them.

Wether they were interested in 3d graphics chips then or not I don't know for sure but according to VIA they are interested now:

“VIA Technologies announced on January 30 the kick-off of its “Canaan Projectâ€￾ to deepen the diversification of its operations. According to the project, VIA will restructure its business into five units – chipsets, processors, graphics chips, optical drives and networking equipment. Its VIA Platform Solutions Division (VPSD) will remain under the chipsets division. Company CEO and president Chen Wen-chi said that VIA will try to position itself as a “platformâ€￾ provider, instead of just being a core logic chipset designer.â€￾

Not at all, its not about wether its easy or not, its just very possible for a big company thats focused on the PC market like VIA to do what I said.

 

Agreed. That would shove them into the integrated chipset ideal, which is never going to get anywhere near high end.

I'm certain this wouldn't happen, it wouldn't exactly be what a IP Licensing company would be looking to do and would probably be considered as a devolutionary step. The only way it could possibly work was if it was bought and merged with the Videologic Systems Division, but like I said, I'm certain that it won’t happen anyway.

What happened to that consortium of companies who were supposed to be working with Sega and ImgTec on a new console? Hitachi and Philips were also included with this. Obviously the console is no longer happening, but these guys co-developing graphics solutions would certainly be interesting. Unlikely, but interesting

 

Wouldn't Intel be a real possibility? They're involved with integrated graphics, mobile graphics and (in the past) AIB. Unfortunately, their 3D tech hasn't been up to much. They could also hardly be accused of not being interested in the high end - how many people buy a 2GHz+ CPU? The biggest concern is that AMD's competition may encourage them to focus on their core business.

 

Said consortium was misconstrued and information on it is not available. SEGA will reenter the home market only if it goes back to $2 billion in black. This probably will happen within 5 years, but don't get your hopes up. Phoenix will be obsolete by then most likely, but it would cost alot less. Does anyone think it would be possible to get NURBS in real time below a price of $15,000 for a unit?

Sonic

 

One could perhaps make a case for IMG developing flagship products themselves, and license their tech for other areas (mobile/integrated/STB etc.)

Their partnerships with NEC and ST have hardly resulted in rapid development of PVR PC chipsets (and hence have had only a limited effect on the market), and this is very possibly because the relationships between IMG + partner are too unwieldy.

There is also the argument that IMG should have a larger say in the development of 3D standards (they are a non-voting participant at OpenGL ARB meetings). With an unorthodox architecture, surely they should be ensuring that 3D APIs develop in a TBR-friendly manner? Of course, in order to influence standards, they must be at the bleeding edge of 3D tech themselves - where they're currently not.

IMG's IP licensing business model seems to be based on ARM's.

ARM is perhaps the most successful, high profile IP licensing company. The mature CPU industry has the following divisions:

a) High performance, high cost, low turnover dominated by the likes of specialist CPUs from SUN and IBM.

b) Medium level desktop CPUs from Intel, AMD, Motorola.

c) Low cost CPUs in commodity devices, particularly SOC, such as phones, PDAs, STBs, dominated by ARM.

The ARM was originally designed by Acorn as a desktop CPU. StrongARM was ARM's last attempt at a 'big' CPU, and good though it was, it never was hugely successful, and is now used almost exclusively in commodity devices.

The less-mature 3D market doesn't have a commodity section (maybe MBX will start to create one). Bracket a) is for the likes of 3D Labs, bracket b) for ATi and nVidia. I suggest the IP licensing model, so successful for ARM, is most appropriate for a bracket c), a type of market which doesn't really exist for 3D.

(Although... could IMG even afford to buy STM's gfx division?)

So were ST Micro . Actually, I'd love to know the answer to this too, but noone seems to know.

 

Sonic: care to spill the beans on 'Phoenix'? Does it use PowerVR graphics? Can we infer from your description of Phoenix the level of performance of which PVR is capable? (IIRC, Sega stated that they would continue to use a combination of Hitachi + PowerVR in future hardware). OK, you probably won't sing, but there's no harm in asking.

 

I think announcement needs to be made soon, IMG are loosing alot of confidence in the financial markets. They need to be seen to be acting fast, and securing their long term future. A deal with VIA or SIS will definately turn things around for the company, but it is strange that a buyers name has not been touted yet....

 

Another french interview of Hercules there
http://www.inpact-hardware.com/read_comment.php3?id_news=4466

I've translated (systran) a quote for english readers :

""""""""

- What about Kyro 3?

Christophe: We are not accustomed to speaking about products which are not announced officially.

- According to some sites, Hercules could consider the purchase of the graphic division of STM. Is it only a false rumour, or on the contrary, is it founded?

Christophe: This rumour is unfounded. On the other hand, WE WILL BE PLEASED TO WORK WITH THE FUTURE PURCHASER OF THE GRAPHIC DIVISION OF STMICROELECTRONICS.

"""""""""

In my opinion

The Kyro 3 isn't announced officially but it will

The future purchaser of STM's graphics division is no more hypothetical.

 

I don't agree.

Even if Hercules would have no plans to sell an card based on Kyro3 it is
very likely that they would give an press-blurb like "we don't comment on
future products" cause otherwise they would only hurt IMG and themeselves.

Also I find it disturbing that they see the Kyro3 as an mainstream-card like
the Kyro2. IMHO this could only mean that the Kyro3 is not as fast as an
Radeon8500 and has not enough features to compete with the Radeon7500 too.
I hope thats not true.

Also; the comment about the future purchaser of the STM-division only
means that Hercules would like to work together with them, but not that
an purchaser is already found.

IMHO, when STM would have had an purchaser of the graphics-division
already they would have only made an press-release saying "this is the
new company producing Kyro-cards in the future, we wish them all the
best". If an purchaser would have been close to be found, then we would
only heard about all this afterwards too (maybe in a few weeks). Only
when the situation is bad and no purchaser is found or near to be found
then we see what we see now ( announcement of an future sell of the
division ), cause STM will loose a lot of experienced engineers with this
announcement and could kill the future of the Kyro-3D cards too.


Manfred

 

we have a comment from David Harold mentioning that even if no one else produces PowerVR cards, Videologic Systems is still committed to the PC space with the POwerVR technology.

see

http://www.pvr-extremist.com

so at least there will be Series 4 and 5 cards out there....timescale...who knows....

 

Why would the Kyro III being a mainstream card mean it doesn't have as many features as the Radeon 7500? The 7500 is basically a faster version of the original Radeon (same features), your saying the features of the original Radeon are beyond mainstream? I don't think so.

AFAICS it'd be hard to see the Kyro III not being at least as fast as the Radeon 8500. Even if the Kyro III was only 200mhz it should still be faster then even the 300mhz Radeon 8500 simply because it'd be a TBR with only 50% less bandwidth then the 8500 and a 800mpixels/s fillrate which it would likely fully hit.

Also when a company says a card is for the mainstream they mean price not necesarilly speed or features.