Larrabee

[Nemesis 1]

Well I will wait and see ,I said what I have to say about larrabee I not try to sell anyone on that . But Haswell I think should be kept in mind along with AVX and how this Fits into Intels long term 3 year outlook. Will look at larrabee than after seeing it in action Haswell becomes very interesting or not so interesting. Those are my thoughts,

 

[IntelUser2000]

Originally posted by: Idontcare

*Omitted*

Very well stated BenSkywalker. My opinion on Larrabee and Intel has slowly evolved over the past 18 months, and I've got to admit that what you state above pretty much reflects my sentiments of it now. (because you wrote it down, now I can see it, and align my thinking to it, very rational, logical, reasonable...thanks!)

Larrabee is a BIG question mark. On one hand, Intel is experimenting with cores they are very familiar with, and that might ease development in both hardware AND software(drivers) being x86.

Other is that whether they can translate that into advantages(mostly performance). And their experience in this area is in infant stages.

Last time similar thing happened(but on a small scale), Nvidia didn't predict well worth a jack. Months before releasing Extreme Graphics aka 845G, Nvidia lowered the prices of their low end graphics products, only to find out that it wasn't even needed. The rumors were something like "they were afraid what Intel might do with the 845G". This is too big to speculate based on rumors alone.

 

[William Gaatjes]

I was just looking around the internet, and i have a question for the people who have knowledge about rendering pipelines and videographics.

Matt Pharr works for neoptica which was purchased by Intel.
He used to work for Nvidia as well and he wrote books about gpgpu computations.

On this link is a pdf where he explains a new theoritical framework for rendering.

scattering rendering

When reading the pdf about the slides from talk, On page 7 i noticed a few big names from the physics field. If i am not mistaken, Intel does a lot of physics research themselves. Intel is bound to have a lot physics knowledge themselves. Is Intel doing research for gpgpu supercomputing on the physics field and is larrabee just a start for a supercomputer filled with larrabee's or future incarnations of larrabee ? gpgpu can be used for many applications and this would be one of them i assume.

Here is my question :

Is it really something new and advanced he is proposing for graphics ?
He is lead architect of the advanced rendering technology group at Intel.