Intel Discrete GPU Roadmap Overview

[zittware]

"Intel's Visual Computing Group (VCG) gave an interesting overview of the discrete graphics plans this week. There seems to be a few interesting developments down the pipeline that could prove quite a challenge to NVIDIA and AMD in 2 years time.

Their first flagship product for games and graphics intensive applications is likely to happen in late 2008-09 timeframe and the GPU is based on multi-core architecture. We heard there could be as many as 16 graphics cores packed into a single die."
-- http://www.vr-zone.com/?i=4605
confirmed here:
"Larrabee project "
"Look out, NVIDIA and AMD/ATI, here we come!"
http://www.beyond3d.com/forum/showthread.php?t=37889


This should quell any rumors to an nVidia / Intel partnership.
I for one welcome this entry - as HOPEFULLY it'll kill nVidia's plans for world SLI domination.
Gawd I hope Intel champions an Open SLI standard.
_________________
Worklog: ATTN2Detail
Worklog: RPM Flow Sensors for Liquid

[PoMan]

Maybe while they are at it they can end world hunger? Stop Global warming? Cure Cancer?

Just picking on your Intel-ishness...I think it WILL be great if they can become another player in the GPU market. But to say they will be ending nvidia's rule ,after reading an [intel written] news brief, is craziness. Sad to say, they got a long way to go from "ExtremeGraphics" to something that tops a real video card On the other side, if any company HAS the resources to use, to get up to speed with the market, it would be Intel. I know they arent starting from scratch, so there is a possibility something will come of it.

Time will tell. BTW, did you know Via makes plug in video cards? they are only like 4 generations behind nvidia/ATI.

Dont take it as "anti-Intel". But to me this is a VERY tough market. And I'm not believing ANY marketing hype until I see hard data...
_________________
No matter how you shake and dance, the last drop always falls in your pants

[zittware]

Agreed.
Intel was in the GPU market breifly with a couple of products about 4-5yrs ago.

I know the market is tough; I'm not sure if the big I can pull it off. As I've stated before; the current big 2 players are all about time to market - not time to quality. Intel has got to drop the "validate the crap" outa the silicon in favor of "ship it and fix it in drivers". It would be a tough sell.

However, the way the current GPU designs are going - I can see where there may be a better opprotunity since GPUs look a lot like specialized CPUs. Technology-wize it might be possible for good designs to appear.

I don't doubt for a minute that nVidia will continue to fight. But maybe just maybe it'll influence another change like what happen 4-5yrs ago.
_________________
Worklog: ATTN2Detail
Worklog: RPM Flow Sensors for Liquid

[PoMan]

agreed...MFG capacity could maybe help bring prices down some more also.
I mean, I COULD see many benifits in having Intel (or ANY new player in the market) because prices are going up. I would LOVE to see a solution for open multi GPU, but this isnt a huge open market. both players already have a multi gpu solution. I doubt either will abandon their tech. for something that is open.
_________________
No matter how you shake and dance, the last drop always falls in your pants

[Lancer]

I will be happy to see intel enter the market. There is not enough competition and the players only have to make enough innovation to keep ahead of each other. I remember when Intel gave it a shot awhile back. Hopefully this was a good lesson for them and they learned a lot. Intel has the cash to be able to do it, but like Zitt said I fear that they do not have the speed. Intel can make some great technology they just tend to not get it out in a very timely manner. Even if Intel's endevor fails hopefully it will push AMD and Nvidia to release some even better products.
_________________

[zittware]

[Lancer]

If Intel wants to seriously compete they will write the drivers, along with making sure they work just as good on that platform as it does their own. The industry will tear them apart if they don't. That will be one of the first tests reviewers put them through. If Intel takes the anti-competitive stance in that area they will be shut down quick. I almost wish Intel would create a seperate company to make the cards but allow them access to Intel proc architecture and fabs and say here is the ball, run with it. But I don't see that ever happening.
_________________

[zittware]

[zittware]

Any of your Graphics Architects by day?

http://www.intel.com/jobs/careers/visualcomputing/index.htm
We're developing advanced products for high-end client platforms and our GPU silicon designs have us aggressively positioned to advance the state of the art in graphics and other high-throughput workloads

We’re looking for engineers, developers, and architects who share our vision and understand what can happen when serious skills and vast resources join forces.
_________________
Worklog: ATTN2Detail
Worklog: RPM Flow Sensors for Liquid

[MrArmyAnt]

Right, 1 cores on a single GPU, much like standard DVD's were theoretically capable of 80gb a disc. Theory it works, in reality they will be no far ahead than the rest.

Go intel, I hope they do some good.
_________________
"If it ain't modded, what's the point?"
~Copyright MrArmyAnt, 2002-2006, or end of time, which ever is later.

MacBook Pro Running Vista - Home of the gender confused laptop.

[zittware]

Anandtech posted some details on
Intel's Larrabee Architecture Disclosure: A Calculated First Move
which I thought was an interested read. Lots of technical discussions which I thought was fairtly well balanced.

Given it's a 2009/10 product... obviously no samples or demos are available.
_________________
Worklog: ATTN2Detail
Worklog: RPM Flow Sensors for Liquid

[zittware]

[Mickey21]

I predicted this so much, I didnt even comment on the thread, surprising actually. FAIL, Intel!
_________________