Nvidia wins $25M defense research grant

[shangshang]

this is interesting to me
http://sanjose.bizjournals.com/sanjose/stories/2010/08/09/daily14.html?ana=yfcpc

I like how NV is involving themselves with the research and academic research universities. It's like free R&D for all practical purposes. I'm sure by working with other high talents in this field, Nv could stand to reap lots of knowledge down the road. Are they pushing CUDA? lol

 

[OCGuy]

nV+DARPA?

Pretty cool.


Wait...didnt they hear about GTX480?? And how nV is going from $900M to $800M in revenue?

Stupid DARPA!

 

[Lonyo]

nV+DARPA?

Pretty cool.


Wait...didnt they hear about GTX480?? And how nV is going from $900M to $800M in revenue?

Stupid DARPA!

Build it and they will come.
$25m in research grants to help them build software and develop technology to make Fermi and its ilk useful (since most stuff isn't yet built for Fermi/GPGPU).

Anyone sane knows that Fermi isn't a now product (in the HPC market at least). They have to start somewhere. They've made a product, now there are people interested in the concept/technologies, and future developments, like writing the software to use the chips (such as this $25m relates to) means their HPC sales can have a bright future.
A decline in revenue now, before there is a real built HPC market, is of little surprise and has nothing much to do with HPC.

 

[shangshang]

I'm thinking for a company of Nv size, in this economic environment, $25M is not something extremely large, yet it's no chump change either. However, with a project like this, Nv scientists get access to all the other bright minds in the field, at no extra cost to Nv. That is an amazing R&D strategy pathway. Say what you want of Jensun, but the guy knows where to go looking for bright minds, go right to the source of the country!

And when the GPGPU field does mature more, and when these scientists from all these research centers and uni research labs go looking for hardware to build their dream computer, they are prolly gonna be favoring Nv a bit. That's how it usually works.

 

[SlowSpyder]

Somewhat related, does anyone know what's going on with that whole Fermi supercomputer? Charlie said it was cancelled, Nvidia said it's not. I'm curious to see who ends up with egg on their face with that one.

*edit - This story: http://www.semiaccurate.com/2009/12/16/oak-ridge-cans-nvidia-based-fermi-supercomputer/

 

[Idontcare]

Somewhat related, does anyone know what's going on with that whole Fermi supercomputer? Charlie said it was cancelled, Nvidia said it's not. I'm curious to see who ends up with egg on their face with that one.

*edit - This story: http://www.semiaccurate.com/2009/12/16/oak-ridge-cans-nvidia-based-fermi-supercomputer/

Charlie can't be sued for lying and spreading false information. Nvidia can be.

Which do you think knows the truth and is telling the truth?

 

[Chiropteran]

Charlie can't be sued for lying and spreading false information. Nvidia can be.

I got a good laugh out of that one. While nVidia might be held responsible for a clear and obvious lie, it's not very hard for them to say something isn't cancelled and later reveal that "well yeah it wasn't canceled at the time of that announcement but later changes in the market deemed that we had to cancel it..."

 

[SlowSpyder]

Charlie can't be sued for lying and spreading false information. Nvidia can be.

Which do you think knows the truth and is telling the truth?

I could be wrong (can't find the link refuting Charlie's claims) but I don't think Nvidia ever went on record when it was denied. I think it was a story where a source from Nvidia said that Charlie is wrong, but I don't know if it was an official on the record statement. I'll see if I can find it.

*edit - Employees at Oak Ridge 'had no idea what we were talking about' when LegitReviews asked them about the cancellation. Someone in the Computing and Computational Sciences said that the article was innaccurate.

So it looks like Charlie was off on this one, but I still want to see if this thing is built or if it just never shows up. I'd like to see Charlie explain this one if it does get built...

http://www.legitreviews.com/news/6999/

 

[Idontcare]

I'd like to see Charlie explain this one if it does get built...

http://www.legitreviews.com/news/6999/

lol are you kidding? If he has so few scruples as too post such FUD you think he'll suddenly find some scruples and post and honest article explaining why?

It would be very easy for him to just ignore it, if pressed on it then it would be equally easy to just fabricate a claim along the lines that the contract had since been renegotiated and the DOE decided to just live with the fact that they needed to build a tokomak fusion reactor to power the Fermi-based supercomputer. Sky is the limit when your craft is hawking bullshit to get page impressions.

 

[ronnn]

Charlie can't be sued for lying and spreading false information. Nvidia can be.

Which do you think knows the truth and is telling the truth?

Likely Charlie, but generally the truth is in that dark void of the middle.

 

[sandorski]

Nvidia is actively looking for alternative and niche Markets. They may not need them, but increasingly it looks like they will need them.

 

[Genx87]

Nvidia is actively looking for alternative and niche Markets. They may not need them, but increasingly it looks like they will need them.

They will need them as the integrated garbage consumes more and more of the discrete market. And the PC gaming world is consumed by console ports built around 4-5 year old hardware. Nvidia knows this which is why they are venturing into these markets.