ATI acquires Bitboys

[Dave Baumann]

The deal will span both ATI and BitBoys graphics solutions. ATI also has further multimedia additions that I suspect wil be integrated into both. ATI's further role is to provide a software development platform that makes these all look as similar to one another as possible.

 

[Xmas]

It is so obvious and nobody can see it. BB was not a backdoor to Nokia, BB is the frontdoor. BB had a right product for nokia and had most likely already won an account that everyone wanted. They got it. Not ATI, not Nvidia not IMG.

If you knew anything, you would have known for long time that IMG is not the winner in this game. It had an opportunity and it screwed it up with lack of execution. Now the big players are coming into the handheld market, and the "cards have been dealt". Smart move from BB to get there and join one of the big ones, not entering into a fistfight with any one of them.

Nvidia tried the backdoor with Hybrid and its software, but with the announcement of Nokia alliance ATI is at this point right where it wanted, with BB's product in the IP side and its own hardware products in the upper end. Surely the strategy can fail, and Nvidia can pull out tricks and win other accounts, but if ATI / BB can deliver, they have the volume market in the handheld graphics. Period.

This thread has more vaporware within it than what BB ever was able to create. Give up guys, you have to admit that BB achieved something you never will. And by the way, so did Hybrid. Great to exit in style!

So if Bitboys is the front door for ATI and Hybrid is the back door NVidia tried, does that mean TI is the chimney through which IMG came in?

 

[finsider]

Time will tell

So if Bitboys is the front door for ATI and Hybrid is the back door NVidia tried, does that mean TI is the chimney through which IMG came in?

Good point. But think about TI's position now as a supplier of chips to Nokia. Will they be able to afford to bet on a different graphics solution from their major customer? At least not exclusively, which means that they will have to play along as well.

I would bet that TI is surely not likely to acquire IMG. I believe IMG will simply fade away as not a serious contender to ATI or Nvidia.

But, as said, only time will tell.

 

[finsider]

The deal will span both ATI and BitBoys graphics solutions. ATI also has further multimedia additions that I suspect wil be integrated into both. ATI's further role is to provide a software development platform that makes these all look as similar to one another as possible.

Totally agree. ATI's portfolio of patents and technologies also gives a peace of mind to BB developers to use whatever best technology there is available in the hardware solutions. This deal works for all, but not really for Falanx or IMG who are really against the wall now.

 

[Xmas]

finsider, that's a bit too much doom & gloom for my taste. Nokia is the biggest cellphone maker, but they're not the only one. And IMG is already there.
Certainly ATI and NVidia aren't going to leave that market anytime soon, with their strong standing in the PC and console markets. But that doesn't mean they're going to be the only players or the dominating ones. Even in the PC market there's a third big one, remember? And it won't be too long before the handheld 3D market takes over in size.

 

[finsider]

For the time being

finsider, that's a bit too much doom & gloom for my taste. Nokia is the biggest cellphone maker, but they're not the only one. And IMG is already there.
Certainly ATI and NVidia aren't going to leave that market anytime soon, with their strong standing in the PC and console markets. But that doesn't mean they're going to be the only players or the dominating ones. Even in the PC market there's a third big one, remember? And it won't be too long before the handheld 3D market takes over in size.

Sure, and you are right. But for the time being big decisions have been made and these ships do not turn around every quarter, as did the PC graphcis card buyers back in the good old days, when the race was on. So my bet is, that ATI has now the driving seat with a company that does make decisions on a great part of the market. Nvidia will definitely not give up the fight and there are other big handset players for it to win, even just because all of them do not wish to follow Nokia. And IMG will try to hang there, but I do believe strongly, that they missed their early opportunity.

Handheld market is so similar to PC market in many respects as to the development of graphics hardware, but yet there are great differences in the business logic. Consumers do not make decisions on the graphics solution at all, it is all about cost efficiency and performance. Once the market takes off, it will be hard to challenge the ones who have started it, unless they fail to perform. And whoever really gets the 100 million + annual market can probably also sell the higher end separate chip solutions, with software compatibility through the product range. This means savings and that is profits for big companies with big volumes. Remember, each dollar saved is 100 million dollars bottom line for such a volume.

Also keep in mind the vector graphics based solutions. Not yet there, but there is a lot of potential in a company like ATI supporting and probably also using BB vector graphics hardware accelerator core in its products. Nobody else has anything like it yet.....

What will most likely make things very difficult is, that the handheld companies have not had a track record of advertising the exact components they use, so we may know ultimately very little about who did the graphics......

 

[Lazy8s]

The market-shaping decision which will define the near term isn't the ATi deal; it was the OMAP2420 arrangement. ATi expects to start getting Nokia's business, though not exclusively, in twelve to eighteen months, yet the 3D Nokias coming now and those of their next generation games platform for 2007 will be PowerVR.

 

[Davros]

Damm this means ill be waiting for ever for my bitboys gfx card

 

[Humus]

Well, if you were waiting at this point you were already waiting in vain. Bitboys had already dropped out of the PC business and gone over to chips for mobile devices a long time ago.

 

[Davros]

shame, it was going to be more powerfull than the voodoo2

 

[Megadrive1988]

mega-charge for ATI.


ATI R&D - engineering fueled by

-ArtX (part of SGI)
-ATI Orlando (part of former Real3D that had GE Aerospace and Martin-Marietta tech)
-BitBoys (Hammer, XBA, Glaze 3D, Pyramid3D, etc tech)

(amoung other less ATI acquisitions)

not to mention the talent that was already within ATI East / and ATI HQ in Canada


R800 could be something amazing (assuming R700 is just a refresh of R600 which is heavily based on Xenos tech which had its roots in the unreleased R400)

and certainly the next-next gen Xbox will benefit from BitBoy's technology which can now be fully implemented thanks to a vastly larger budget and resources.


sadly, PowerVR has just become somewhat less important. even though I like what Videologic did with the first two generatioms of 3D accelerators during the PRE-GPU area for consumers.

 

[rwolf]

sadly, PowerVR has just become somewhat less important. even though I like what Videologic did with the first two generatioms of 3D accelerators during the PRE-GPU area for consumers.

I wonder who is buying PowerVR.

 

[Lazy8s]

Tens of millions of people with cellphones are who's buying PowerVR. With that rapidly climbing revenue stream from the royalties, Imagination Technologies is building the resources to stand competitive in the market.

This deal consolidates opposition to PowerVR in the sector but also results in one less competitor to it.

 

[sethk]

mega-charge for ATI.

...and certainly the next-next gen Xbox will benefit from BitBoy's technology which can now be fully implemented thanks to a vastly larger budget and resources.

Wasn't BitBoy's technology primarily a high speed embedded RAM/Cache with very high speed local access from the GPU to the RAM/Cache? Isn't that technology already used on the Xenos? And actually on the PS2 as well?
I don't remember what else BitBoys planned to do that still stands out as unique / unimplemented technology.

 

[DudeMiester]

You know as GPUs become more generic, with unified shaders and such, I could imagine implementing them might actually become simpler. There's less types of units, although more total units. That property is common in networking chipsets, which afaik usually have smaller teams then ATI or NV. Perhaps we'll see more GPU providers, like we have with CPU providers, where the major differntiation is not functionality but clockspeed, process , power and efficency.

 

[KimB]

You know as GPUs become more generic, with unified shaders and such, I could imagine implementing them might actually become simpler. There's less types of units, although more total units. That property is common in networking chipsets, which afaik usually have smaller teams then ATI or NV. Perhaps we'll see more GPU providers, like we have with CPU providers, where the major differntiation is not functionality but clockspeed, process , power and efficency.

I doubt it. If the hardware becomes simpler, that just shifts the engineering burden over to the software side of things.

 

[Gubbi]

I doubt it. If the hardware becomes simpler, that just shifts the engineering burden over to the software side of things.

True. But maintaining complexity is generally a lot cheaper in software than in hardware. Software is also a lot more flexible.

The succes of such a move is of course dependent on how much performance you have to give up.

Cheers

 

[hkultala]

I don't remember what else BitBoys planned to do that still stands out as unique / unimplemented technology.

they had a new AA method ("matrix AA"), though this was not used in their mobile IP cores, which use ericssons FLIPQUAD instead.

Matrix AA has some same basic principles as xenos's AA but the implementation is somewhat different. (and the reason it was not used in mobile cores was propably it's complexity and maybe also fast ram(practically edram) requirements.