INTEL Larrabee and threat to AMD/ATI/NVIDIA

[Rockoon]

NVidia is firmy denying rumors of them entering the x86 market..

http://www.pcpro.co.uk/news/221229/nvidia-we-wont-build-a-cpu.html

[Mark_Larson]

NVidia is firmy denying rumors of them entering the x86 market..

http://www.pcpro.co.uk/news/221229/nvidia-we-wont-build-a-cpu.html

every big company does that.  Until they are ready announce it, they firmly deny it.  When big companies deny they are going to do something, take it with a grain of salt.  I've been in the industry quite a while, and this is standard practice.

they have to do it to remain competitive with Intel and AMD/ATI. 

Mark

[Rockoon]

In this case tho, they didnt just say that they 'wont' try to compete with Intel, they are saying they 'cant' compete with Intel.

[Mark_Larson]

In this case tho, they didnt just say that they 'wont' try to compete with Intel, they are saying they 'cant' compete with Intel.

still more garbage.  When they entered the 3d market, they were competing against ATI and 3dfx who had been in the market longer and had more resources and experience.  But yet they still competed and eventually ended up on top.  They eventually bought 3dfx.

I have one friend who is at Nvidia, who I used to work at with at Dell.  He works in the video BIOS group.  He was the one who told me.  Nviidia itself can't enter the x86 market and be competitve.  It would take them years to get to that point.  That is why they have to buy a company with lots of x86 experience.  So technically their statement is true.  It'll be some company that they buy that does it. 

EDIT:
so let's look at two of the quotes from the article. The below quote  isn't true.  They actually sell a range of products including motherboards.  He is just trying to misdirect.  How is a motherboard a "visual product"?  Here is their shopping page.
http://store.nvidia.com

"That's not our business," he insisted. "It's not our business to build a CPU. We're a visual computing company, and I think the reason we've survived the other 35 companies who were making graphics at the start is that we've stayed focused."

This is also completely a misdirect.  The 3d graphics market is just as fierce as the x86 market.  That is why most of the 3d companies went bye bye and only 2 remain.  And again, they wouldn't be doing it.  They would buy a company with many years of experience.  Not to mention the fact that the product cycle with GPUs started at 2years, and kept getting shorter and shorter.  It's about 6 months now.  Do you think Intel could produce a new x86 processor in 6 months?  hell no.  So I would actually argue that the 3d graphics market is even more intense.  In addition have you ever looked at Cuda?  It's a program for programming a Nvidia  GPU directly in C.  I point you to this Dr Dobb's article about it.  They guy doing the article shows you how to turn your computer into a super computer.  He has access to a super computer.  He shows the speed up that is usually in the 50 times faster range for using Cuda over standard applications that a super computer would do.  He has access to one and posts his timings for both.  I highly recommend ya'll read it.  He goes into the times faster on the first page.
http://www.ddj.com/hpc-high-performance-computing/207200659

"He also pointed out that such a move would expose the company to fierce competition. "Are we likely to build a CPU and take out Intel?" he asked."I don't think so, given their thirty-year head start and billions and billions of dollars invested in it. I think staying focused is our best strategy."

everyone agees that the future of computing is to combine a CPU and a GPU.  That is why Intel is doing Larrabee, so they can extend their knowledge in the graphics market.  Currently their graphics solutions have been bad.  So now with Larrabee they are taking it seriously. 

[Rockoon]

still more garbage.  When they entered the 3d market, they were competing against ATI and 3dfx who had been in the market longer and had more resources and experience. 

Note the differences, not the similarities.

a) The 3d market was rapidly growing.
b) They didnt use their own money. Instead Sequoia Capital took the risk, lucky for them

Do you think that before they asked Sequoia for money, they announced that they couldn't compete with ATI or 3DFX, and then went shopping for money anyways?

Here, they announce that they cannot compete with Intel or AMD, pretty much putting the nix on ever getting any external investors. Also rules out mergers, stock trades, and other common methods of aquiring another companies assets. Their only choice now it a flat buy-out, and I highly doubt that they have enough liquidity to do so on their own. They can't borrow the money now, they cant use their stock as collateral now, and a massive public offering would open up the door for AMD or Intel to gobble them up.

..and even after they aquire a license to produce x86 clones, they still have to spend a billion or two on R&D, since nobody else has any competitive technology which they could aquire.

IMHO NVidia is highly unlikely to be making x86 clones any time soon, if ever. They will eventualy downsize.

[hutch--]

Mark,

The CUDA stuff looks interesting, I wonder how general purpose this stuff is capable of, the massive multithreading interests me.

[Mark_Larson]

Mark,

The CUDA stuff looks interesting, I wonder how general purpose this stuff is capable of, the massive multithreading interests me.

extremely general purpose and highly parallel, that is what shocked me.  I"ve seen raytracers using it.  Pi programs could use it, because they have a FFT.  I have it installed under Linux and I am playing with it, but I am having a problem compiling code.

For the geeks, I recommend ya'll download it and play around with it.  It is way cool :)

[bozo]

i'm running it on windows and it is worth learning how to use even if you don't use it for an everyday purpose.
i have a feelin Larrabee will offer more power though..we'll have to wait and see.