NVIDIA shows signs ... [2008 - 2017]

[Silent_Buddha]

Ouch on losing that sale to China. ~Half a billion USD is a big chunk of change. It'll be interesting to see if it's a one time payment or if it'll be amortised over multiple quarters for AMD.

[edit] - changed 500 billion to half a billion. Ooops.

Regards,
SB

 

[AlexV]

Ouch on losing that sale to China. ~500 billion USD is a big chunk of change. It'll be interesting to see if it's a one time payment or if it'll be amortised over multiple quarters for AMD.

Regards,
SB

Ugh, you're been spending too much time with Dr. Evil, and billions are the consequence...The US would cry upon losing 500 bln USD, luckily NV only lost (supposedly and quite unconfirmedly) 500 mln USD. Maybe they value their IP or something...

 

[silent_guy]

Ugh, you're been spending too much time with Dr. Evil, and billions are the consequence...The US would cry upon losing 500 bln USD, luckily NV only lost (supposedly and quite unconfirmedly) 500 mln USD. Maybe they value their IP or something...

Am I the only one who thinks that releasing the source code of your crown jewels to the Chinese is a really bad idea? I wouldn't be surprised if that would be disallowed under various trade restrictions. (I suppose that's what you meant with valuing their IP?)

 

[Sxotty]

Am I the only one who thinks that releasing the source code of your crown jewels to the Chinese is a really bad idea? I wouldn't be surprised if that would be disallowed under various trade restrictions. (I suppose that's what you meant with valuing their IP?)

Yeah companies that are doing that are selling the future for a short term bump in profit. Glad nvidia was smart enough not to.

 

[Kaotik]

Yeah companies that are doing that are selling the future for a short term bump in profit. Glad nvidia was smart enough not to.

...right, because AMD having had open source drivers for anyone to look at for years clearly sold their future?

 

[Alexko]

Ugh, you're been spending too much time with Dr. Evil, and billions are the consequence...The US would cry upon losing 500 bln USD, luckily NV only lost (supposedly and quite unconfirmedly) 500 mln USD. Maybe they value their IP or something...

That's a very big contract. If there's any truth to this report, I would imagine that AMD will make a press release, or at least mention it at some earnings conference.

If they don't, we can assume it's bullocks.

 

[Blazkowicz]

The problem is that the GeForce / Quadro driver from NVIDIA is only available for Linux x86 and x86_64 architectures, not MIPS or even ARM (only the Tegra driver is for ARMv7). NVIDIA refused to release the source-code to their high-performance feature-complete cross-platform driver to the Chinese, and it would cost them millions of dollars to port the code-base, so they went to AMD for their GPU order.

when I go to nvidia's website, download section, there's no option for Tegra drivers, only geforce, quadro etc.
you know, the full linux driver thing. with Xorg, OpenGL, vdpau etc. , I guess there's very little market for that, someone who installs full blown GNU/Linux on a Tegra tablet/laptop "transformer", but hell why not.they support FreeBSD, that's enough to say.

They'll have to port to ARM, eventually! else their Denver APU would fail. they'll support x86, x86-64 and ARMv8. but they don't seem kin on lower ARM, MIPS or something else.

the Chinese should have paid. but they are after low-cost licensing, and controlling their MIPS-based stack. why not, nvidia couldn't meet such demands so they made a big fuss about it to make the blame rest on nvidia.

 

[Ryan Smith]

They'll have to port to ARM, eventually! else their Denver APU would fail. they'll support x86, x86-64 and ARMv8. but they don't seem kin on lower ARM, MIPS or something else.

So far all NVIDIA has seriously talked about using Denver for is as the master CPU for Tesla clusters. If that's the case they merely need to port their Tesla computer cluster drivers, which is far easier than porting the display driver.

 

[AlexV]

(I suppose that's what you meant with valuing their IP?)

Yes. Also (in reply to Kaotik), whilst it's cute and all Xorg Radeon or whatever it's called is not fglrx. There is a very significant difference between the two. If the existence of an open-source driver developed by 3 gentlemen and a few not necessarily great documents was the discriminant, then I bet even a low ranking NV guy could've pointed out Nouveau.

My humble assumption is that you need a slightly more significant nudge to take your money one way or the other, so if the decision is true it was likely motivated by something else. The "oh noes, lack of FOSS as Linus said is killing NV's business" meme is silly. There are very expensive, Chinese supers built using Teslas...think they got source-code access to NV's main trunk for those?

 

[silent_guy]

So far all NVIDIA has seriously talked about using Denver for is as the master CPU for Tesla clusters. If that's the case they merely need to port their Tesla computer cluster drivers, which is far easier than porting the display driver.

There was this tidbit some time ago: http://www.xbitlabs.com/news/other/...ing_of_Supercomputer_App_Development_Kit.html
It doesn't say anything about the display driver, but you'd expect software of this kind to be written with a respectable amount of abstraction layers. If they can control a Fermi chip from ARM for compute, you'd think the hardest part has been done?

 

[Blazkowicz]

I thought Denver would be a fit for the "from tablets to supercomputers" meme.
what we know is nvidia only talks about compute / GPGPU when mentioning long term, medium term products.
or even any unreleased product, such as gk110.

 

[Silent_Buddha]

Ugh, you're been spending too much time with Dr. Evil, and billions are the consequence...The US would cry upon losing 500 bln USD, luckily NV only lost (supposedly and quite unconfirmedly) 500 mln USD. Maybe they value their IP or something...

Whoops, my bad, I meant to type "half a billion" when I was writing "500 million" and ended up typing "500 billion" instead. /mud_on_face.

And half a billion is still a large chunk of change for what amounts to a single transaction even if it is amortised over multiple quarters. That's basically over half of the revenue of Nvidia's last financial quarter.

I'm assuming, and this could be wrong obviously, that the Chinese are getting the same or similar access to the AMD source code base as the open source Linux guys are. Which I believe is everything not covered by patents owned by other companies or licensed from other companies.

Regards,
SB

 

[madyasiwi]

when I go to nvidia's website, download section, there's no option for Tegra drivers, only geforce, quadro etc.
you know, the full linux driver thing. with Xorg, OpenGL, vdpau etc.

This isn't meant for end user.

 

[DSC]

http://www.phoronix.com/scan.php?page=news_item&px=MTEyOTU

http://www.phoronix.com/scan.php?page=news_item&px=MTEyOTI

As China will soon find out, AMD's crappy closed or open source drivers will cost them more in the long run.

Also, Nvidia already won big design wins in Macbook Pro Retina, Surface tablet and Nexus 7 tablet.

 

[Kaotik]

Reportedly nVidia has had to push Tegra price down quite a lot, not a surprise either since others are bringing Cortex A15 and Qualcomm brought their Krait.

The crappy drivers is getting tiresome already, even if they're not nV level in Linux

 

[Davros]

amortised

Spread out

 

[Davros]

NVIDIA Hacked, Usernames & Emails Stolen

er day, another hack of a large online community, this time PC graphics specialists NVIDIA. The company's forums were hacked last week, with culprits making off with usernames, email addresses, public information in their profiles and "hashed passwords with random salt value".

NVIDIA has closed its forums down while it investigates, and is urging registered members to change their passwords on other sites (if they shared a password).

 

[liolio]

Good quarter for Nvidia:
http://techreport.com/news/23872/nvidia-posts-record-quarterly-revenue

 

[fellix]

NVIDIA GPU marketshare increases as AMD and Intel drop

 

[UniversalTruth]

Nvidia Tips 20nm Contract to TSMC

http://news.cens.com/cens/html/en/news/news_inner_42252.html

Eisler said although his company has evaluated all foundry options, TSMC remains the best choice for his company. He noted that his company has set up long-term cooperation with TSMC, which has persistently provided Nvidia the optimum portfolio of production capacity, technology and pricing.

Samsung Electronics is also reportedly competing for Nvidia’s 20nm foundry contracts.

Samsung’s share of world silicon foundry market remains low. According to market research organization Gartner, in 2011 TSMC had 48.8% of the market, far leading United Microelectronics Corp.’s 12.1%, GlobalFoundries Inc.’s 12% and Samsung’s 1.6%.