Intel larrabee pics

[nforce4max]

 

[wa1]

wow...
Could you give more info/link for these things, nforce4max?

 

[nforce4max]

Search in google images Intel larrabee and Intel Knight's furry. It will take some effort to search through the spam but you will find it. There isn't much out there for me to link :/

There are slides as well hinting at it's performance. Max is just over 520gflop. I got more shots. It is a 256bit if I am right but it has 8 vram modules with 4 oddly placed.

 

[jprahman]

530gflop? That's pretty weak. The max single precision floating point performance for each of my GTX 460s is about 1000gflops, and GTX 580s run up to ~1400 single precision gflops. Looks like this going to be targeted at HPC mainly anyway, it gets blown out of the water by current gen video cards.

 

[plznote]

I though Intel abandoned Larrabee?

 

[jprahman]

They did abandon it as a consumer graphics card. However, they decided that they didn't want to waste all the R & D money and effort, so they are planning on releasing it an academic research / High Performance Computing (HPC) product.

 

[nforce4max]

I though Intel abandoned Larrabee?

Nope they are going to bring a few to market using the 22nm production next year from their own fabs. I think that is why they even bothered to spend the money in the first place in R&D just to make it from them selves. 32 cores at 1.2ghz. was 48 but articles from around 2009 suggested that it had many problems.

 

[nforce4max]

http://software.intel.com/en-us/blogs/2011/02/09/intel-gaming-and-graphics-samples-blog-post-1-introduction/?wapkw=%28larrabee%29

http://www.pcgameshardware.com/&menu=browser&entity_id=154104&image_id=1009150&article_id=681498&page=1

 

[Randomacts]

Intel + Graphics = it prob sucks

 

[christop]

I bet if they really tried they could make a super bad ass gpu.

 

[jprahman]

I'm sure if they threw enough money at it they could given the engineering talent they have. Unfortunately they're stuck in the mindset of avoiding fixed function hardware and they're also not willing to give up programming flexibility in order to get more performance. Nvidia and AMD are able to get so many execution units (and thus gflops) on their GPUs because the execution units are primitive and use a simple instruction set. Intel is trying to use x86 cores, and because of the greater complexity of the x86 instruction set they aren't able get the same performance with the same die size, although it does allow for greater flexibility.