Larrabee: A fiasco, or the future?

[Periander6]

Rumormongers seem to paint very different pictures:

http://vr-zone.com/forums/485017/pat...e-fiasco-.html

http://www.semiaccurate.com./2009/09...out-weeks-ago/

What do we actually know? Is there any solid info out there? Any published roadmaps, specs, or anything similar?

[perkam]

Larrabee = R600/G92 levels at best.

Perkam

[bowman]

bsn is full of and theo valich is someone who posts fiction along with news, way more so than charlie.

I'd have a 'wait and see' attitude to this. It's not the first time Pat Gelsinger has quit Intel anyway.

[eleeter]

I'd have a 'wait and see' attitude to this.

You're in for a long wait.

[Sh1tyMcGee]

Larrabee is kind of like Lucid Hydra, Blink and its gone in the wind!

[nn_step]

in terms of good design : Fiasco

in terms of feeding off your old successes: Intel's Future

[kl0012]

Future, I think. One way or another one day all GPUs become more general purpose.

[BeepBeep2]

Fiasco, I can't imaging an intel graphics card, it may turn out bad...ATI+Nvidia is enough imo

[saaya]

Larrabee = R600/G92 levels at best.
Perkam

and you base that on what exactly?

the first silicon was broken, so what?
did anybody honestly expect intels first cpu-gpu merger will work right from the start?

[SinAugustus]

its only going to be the "future" because intel is attempting to throw the market in that direction

[Dimitriman]

My bet is Larrabee goes the way of the Duke Nuken...

[turbox997]

It's only going to be the "future" because it ain't here today. =P

I admit I don't know much of larrabee, but raytracing looks pretty promising.

[saaya]

My bet is Larrabee goes the way of the Duke Nuken...

no way, they will release it even if it sucks...

if they dont manage to graphics back on the cpu and to x86, then it means gpus will manage to get cpu functionality merged in... and that means, long term, bye bye x86 and bye bye intel... intel NEEDS larrabee...

[AMDDeathstar]

You would have sneak into Dreamworks and Pixar to find how well Larrabee is working.

Think more renderfarm then gaming.

From Motley Fool Sept. 12

And if today's action wasn't hot enough for you, keep an eye out for Intel's Larrabee GPU chips next year. Intel hasn't been a force in high-performance graphics hardware since the early 1990s, but that should change with Larrabee. DreamWorks Animation (Nasdaq: DWA) CEO Jeffrey Katzenberg waxes poetic over the product, saying that "Larrabee raises the bar of what we can do not just by 2X or 3X but by 20X."/

http://www.fool.com/investing/genera...-mountain.aspx

[Martijn]

I think Larrabee will be a lot like the Intel 740 graphics chip... Lots of talking during development, great performance at the time, but outdated by the time it'll be released...

[villa1n]

It's not a matter of If it will pan out, but rather when. How much money and how much ROI is what will dictate whether the product comes to fruition or not.. you cant tell me a company with 10 times the market cap of its top 2 competitors combined cant afford to hire engineers, developers to come up with something better than what is out there...its just a matter of resources.

[FischOderAal]

It's not a matter of If it will pan out, but rather when. How much money and how much ROI is what will dictate whether the product comes to fruition or not.. you cant tell me a company with 10 times the market cap of its top 2 competitors combined cant afford to hire engineers, developers to come up with something better than what is out there...its just a matter of resources.

Well, first I have to say that I have basically no knowledge about ICs whatsoever. But I don't think you can "easily" start from scratch and release a GPU that is as fast as the ones of your competitor, which have been in the business for several years. Somehow I doubt it's that easy, even with a f*ckload of resources.

[Calmatory]

Intel does not adapt to market. Market adapts to Intel.

Thus said, no matter how Larrabee turns out, Intel will make sure the market will follow it's lead.

[Helloworld_98]

if it's aimed at the GPGPU market, it'll do well

if ray-traced games are around when it comes out, it'll do well

otherwise, I don't think it'll do that well.

[DelaMaris]

Would intel release low performing chip that is outdated by today's standards? I don't think so...

[Syn.]

Would intel release low performing chip that is outdated by today's standards? I don't think so...

Yea, its not like they never done that before

[crystogod]

I think larrabee without getting too political about it all is Intel saying "GTFO gpgpu" as the more your offloading to the GPU the less important big expensive processors become.

Intel wants to ensure that their product is the main part that determines performance and at the moment it looks like they will be taking a backseat in the next couple of years if the developers start moving towards largley multi-threaded and gpgpu applications.

Nvidia -> Processing >> Graphics Core
Intel -> Graphics >> Processing Core

A tugg-o-war it seems

[kl0012]

if it's aimed at the GPGPU market, it'll do well

if ray-traced games are around when it comes out, it'll do well

otherwise, I don't think it'll do that well.

With the properly implemented driver the software rasterizer my yield very good result on Larrabee. Here is interesting article:
http://software.intel.com/en-us/arti...n-on-larrabee/

And with that, we conclude our lightning tour of the Larrabee rasterization approach, and our examination of how vector programming can be applied to a semi-parallel task. As I mentioned earlier, software rasterization will never match dedicated hardware peak performance and power efficiency for a given area of silicon, but so far it's proven to be efficient enough. It also has a significant advantage, which is that because it uses general purpose cores, the same resources that are used for rasterization can be used for other purposes at other times, and vice-versa. As Tom Forsyth puts it, because the whole chip is programmable, we can effectively bring more square millimeters to bear on any specific task as needed - up to and including the whole chip; in other words, the pipeline can dynamically reconfigure its processing resources as the rendering workload changes. If we get a heavy rasterization load, we can have all the cores working on it, if necessary; it wouldn't be the most efficient rasterizer per square millimeter, but it would be one heck of a lot of square millimeters of rasterizer, all doing what was most important at that moment, in contrast to a traditional graphics chip with a hardware rasterizer, where most of the circuitry would be idle when there was a heavy rasterization load. A little while later, when the load switches to shading, the whole Larrabee chip can become a shader if necessary. Software simply brings a whole different set of strengths and weaknesses to the table.

More about Larrabee here:
http://software.intel.com/en-us/articles/larrabee/

[ajaidev]

I think Future, only if Intel put's lots of cash behind developers to come up with special raytracing games.... raytracing is the only and the best feature of Larrabee, Intel has the money to make some thing like RGR *Realistic Game Rendering*, etc to promote Larrabee like Nvidia did with their TWIMTBP program.

The bottom line is raytracing would be a big big thing, but it is not so in the present. Intel should release some re-releases of some epic games and some new games using raytracing and people will flock to get the card that does not mean that it should suck in rasterization but lets be honest rasterization is ATi's/Nvidia thing, x86 muscles dont get you far much "Emulation is Emulation".

Intel's card also has to up the performance in DPP and SPP, AMD's 5800 and Nvidia's GT300 are suppose to exceed the initial speculative figs from intels card in terms of performance in both DPP and SPP. But x86 arc does scale very well, just add some cores and up the speed and you are done :>

[Hornet331]

if it's aimed at the GPGPU market, it'll do well

if ray-traced games are around when it comes out, it'll do well

otherwise, I don't think it'll do that well.

Yup, think the same. The first iteration will mainly be marketed to the industry and not the consumers.

Btw the vr-zone article comes not from Vr-zone, but form BSN.... *cough* Theo Valich *cough*