NVIDIA's Little CoC

FrgMstr

Just Plain Mean
Staff member
2FA
Joined
May 18, 1997
Messages
58,686
Jen-Hsun Huang promises instant net access, high-def media, video encoding on the fly, hours of usability and a ton of other things with NVIDIA’s “Computer on a Chip” that NVIDIA showed off working in California about 15 minutes ago playing an HD video. Navigation devices and media players design wins are already a reality with products coming soon.
 
Jim Cramer about ten minutes ago:
NVIDIA NVDA: "No, I've been negative on this and Sandisk SNDK for about a year now. No, it's no good."
I wonder how their stock will fare tomorrow. This looks quite promising.
 
Might as well pick up on this in this thread as well.
1207872502qhBtnLrxx2_1_1.jpg
 
So it's going to be a fight between NVIDIA and ATI for the mobile phone integrated chip designs, since Qualcomm's highly integrated ARM based chipsets use IMAGEON technology from ATI. (MSM7200 and MSM7500) Unfortunately I can't link to the technical pages on Qualcomm's website because they took the info down due to their lawsuit with Broadcom. :-< Here's a dinky, super old, press release:
http://www.qualcomm.com/press/releases/2004/040322_ati.html#top


I have no clue what it is. They did not mention anything by name. Pretty shallow on the details.
 
Intel is waiting for the right time to unleash Skynet. On a more serious note, hopefully this inspires Intel to have a spring in it's step.
 
Nvidia have done just about nothing at all for almost two years. Who knows what surprises they have under their hats?
 
Nvidia have done just about nothing at all for almost two years. Who knows what surprises they have under their hats?


Just about Nothing?!?!? They have purchased and allied with several key tech companies and invested BILLIONS in research.
 
"If we can keep rehasing this same core until the year 2012 we can make this much money, because no development costs is golden!"
 
I guess the only company that seems to have the best grasp on this is AMD. But then again i don't keep up that much w/ Intel to really know that for sure ( I am somewhat unfamiliar w/ ntel's next gen processors.. though it's suppose to have an on chip memory controller. I don't konw if it's gonna have other features as well.) I do know that AMD is on the losing end of both the CPU and GPU front, but could pull away with it all by being the first or the best; preferably both; at combining CPU and GPU power to make a better, dynamic platform.
 
"If we can keep rehasing this same core until the year 2012 we can make this much money, because no development costs is golden!"

I hope you realize how illogical it is to use a lack of released new products as evidence of lower investment and R&D.
 
Almost any proc today is more than enough to run what any normal user will throw at it. You combine that with a low power GPU that can run some crazy GFX...... YouTube mobile HD can become a reality in just a few years.

 
The only thing I will find hard to chew is the whole instant net / media access stuff. ISPs are a roadblock these days.
 
Just about Nothing?!?!? They have purchased and allied with several key tech companies and invested BILLIONS in research.

No, I'm not saying they haven't been doing anything. We just have no idea what they're doing. For almost two years all we have seen from them are some minor GPU revisions and another chipset. That's nothing. With all their time and money they've got to have a bunch of big stuff behind the curtains that we know nothing of.
 
Make PC's into consoles... then your developers won't flock to the more stable environment.
 
It seriously needs to say that on his tombstone or something, I had no idea Jen-Hsun was so awesome.
 
I've found the little EeePC running Ubuntu to be usable for most everything I do except video encoding. Hope these concept pieces work and become something.
 
...from the Nvidia tug boat over to the Intel Destroyer.

Maybe Nvidia is buying into their own SLI hype. They plan to stop innovating and just have everyone buy 3 or 4 GPU's for their PC while Intel continues to make their single chips faster. That would explain the ratios in the chart. Flawless execution on that first part so far, now they just need to convince everyone to buy into Quad/3x/Hybrid SLI :rolleyes:

Those my Friend are Bollocks, Intel isn't making faster processors. In fact they are making more processors. Or havent you heard of Larabee and Nehalem?

And then there is these:
http://www.garagegames.com/blogs/59817/11917
http://www.gamasutra.com/features/20060906/monkkonen_01.shtml

Although an old post I think is valuable.

Where are Intel's Single chips right now?
I dont see any
 
Prepare for a titanic battle:

Purpose Built parallel GPUs.

vs.

General Purpose parallel CPUs.

It may be interesting to see if we are reaching a point where the general purpose CPU is enough to handle whatever we need for a while. Perhaps Nvidia will walk away with the prize by betting on graphics being the focus of the bulk of future PCs. They may very well be correct.
 
Long live NVIDIA! I like their style, put it right to the 800lb gorilla that is intel lol. Now lets see what pans out in the next few year, either they will excel or they will get smashed in the face by that gorilla. :p
 
Those my Friend are Bollocks, Intel isn't making faster processors. In fact they are making more processors. Or havent you heard of Larabee and Nehalem?

And then there is these:
http://www.garagegames.com/blogs/59817/11917
http://www.gamasutra.com/features/20060906/monkkonen_01.shtml

Although an old post I think is valuable.

Where are Intel's Single chips right now?
I dont see any

The reason they started shifting to more cores instead of higher megahertz is because they were running into diminishing returns on higher clock speeds. They found that will multiple cores sharing a single task, or several tasks being divided between different cores performance could be enhanced dramatically. They found a real world benefit by changing their approach. Unlike nVidia throwing multiple GPUs in SLI at us, however, computers didn't start freezing up or lacking stability due solely to the extra cores. Having extra processor cores doesn't disable running dual monitors like SLI does. Having multiple cores in processors supplies a real world benefit with virtually no setbacks, including price. Look at the price of a dual core allendale now, vs. single core top performers of the past.

Games are finally catching up software wise and driver wise and scaling better with SLI, but the transition from single core processors provided much larger benefits with much smaller drawbacks than this long and arduous transition from single GPU to multi-GPU that we have had to deal with. Only recently has SLI become very stable and very practical as a high end solution. There has never been a time in history where you could chain together two cards for $300 in SLI (the 9600GT's) and outperform a $700 card only a year old with extremely few stability, heat, and power usage issues. nVidia is pushing multi GPU's big time. To their credit, they are improving drivers a lot, and scaling has improved a lot as well across multi GPUs but multi-core processors are WAY ahead and the benefits are much more tangible, especially in native multi-core apps like Vista, Adobe programs, many newer games, and a lot of other applications. Vista can even assign single apps to certain cores in order to free up cores you need for more demanding applications, and changing affinity doesn't BSOD your computer like turning on/off SLI occasionally does. Let me slap down part of another post I recently made:

Things appear to be heating up a tad between Intel and nVidia, if you follow Daily Times reports & such.... It appears to me that another ulterior motive for nVidia to not release a fanfreakingtastic single card solution right now is not merely because of ATI's lack of superior high end offerings, but geared towards being "Anti-Intel-Chipset". nVidia has been pushing SLI, and tri SLI, and QUAD SLI so hardcore right now ... They don't want you to buy an Intel board with an intel chipset. They want the best enthusiast solutions to require an nVidia chipset. Look at the 9600GTs! They will beat an ultra, but will it run on an Intel board other than Skulltrail? And are they going to sell enough Skulltrails at Intel to seriously compete with nVidia's lower priced SLI offerings? NO WAY!

The 9800GX2 is the only super high end single card solution, and thus the ONLY top performer that will work on the most popular and common Intel chipsets out there. And what does nVidia charge for it? $160 MORE than what you would pay for two 8800GTS 512 cards that will outperform the single GX2! Imagine how much margin they make on those, hence a win for them money wise. Those who really want top top performance per dollar spent are going to actually spend less or at most about the same by buying two 8800GTS's or GT's AND an SLI motherboard, another WIN for nVidia. And for those who already have nVidia boards, are they going to buy ATI cards?? Not likely, which indicates another win for them. nVidia is using their considerable GPU influence they have right now to push their chipsets as well. Look at all the fantastic, cheap, and efficient midrange solutions out there. Look at how much better scaling is now than before. Think they're doing this by accident? If we buy into it hook, line and sinker (like I have so far!) we are putting the icing on their cake!

Anyways, that's my two cents worth. I see nVidia as having motivation to push their multi-GPU platforms more and more. How much more money does nVidia make when they sell you TWO graphics cards plus an nVidia motherboard vs. selling only one graphics card to someone who purchases a board with an intel/amd chipset? It's simple business. They will push whatever they think will make them the most money, like any smart business would do.
 
Prepare for a titanic battle:

Purpose Built parallel GPUs.

vs.

General Purpose parallel CPUs.

It may be interesting to see if we are reaching a point where the general purpose CPU is enough to handle whatever we need for a while. Perhaps Nvidia will walk away with the prize by betting on graphics being the focus of the bulk of future PCs. They may very well be correct.

Nvidia is basically siding with Google on the idea of cloud computing being the future, then the only thing that your computer will need to do is render/process the data and do minor general computing tasks such as managing data streams. This paradigm will be more and more visible if we get better network infrastructure in the US, but you're already going to start seeing it in your homes with media link systems being built into TVs.

Nvidia and many other companies see a world where your devices all connect to central servers (perhaps you buy one for your home network for specialized programs, but, more generally, server farms) that then send all the data to visual rendering units that are cheaper to produce and more versatile. Then you might see a family of 4 with 1 very powerful processing unit tied into their house (this is where you see Microsoft going with the digital home) and several (probably more than 4) rendering units.

The interesting thing here is that both camps are doing exactly what they need to do to remain relevant; both Intel and Nvidia are on the right track. So this is, really, in no way a "shot over Intel's bow."
 
Jen-Hsun did pound on extremely fast video encoding being done on NVIDIA's GPUs. The technology was promised to be in our hands on August of this year. This might sound familiar as AMD showed this off on ATI GPUs in November of last year.

Lets hope this is nothing like AMD/ATI's solution from last year, because although it certainly was one of the fastest encoders out there, it managed said speed, by having the worst possible quality of any H264 encoders i had ever seen and my a long margin. It also didnt use the GPU at all, as there was various hacks that showed it runnin on motherboards with intergrated intel chipsets.
 
I wouldn't mind seeing a Geforce HD99999GTX-UltraX2 integrated in every Nvidia chipset with Hybrid Quad SLI+CrossfireX.
 
When its all said and done, we all should be absoultly excited about the possibility of a technology and graphics "arms race" between Intel and nVidia. This can only mean GREAT things for the consumer, and even better things is AMD gets in the mix too.
 
diminishing returns in speeds will lead to smarter cpus. a gpu is still a cpu style chip... just with registry aimed for graphics.

we could have a ppu... for physics/math... gpu for graphics... and cpu for other/o/s etc...

then... if parallel processing....
the gpu can handle multimedia, browsing...any media (games/music)...
cpu handle misc
ppu phyiscs/math... (maybe part of the gpu).. we still need a math type chip

blah blah... it makes sense... but they need better tech
 
The only thing I will find hard to chew is the whole instant net / media access stuff. ISPs are a roadblock these days.


There's going to be mass changes about that in the future. Now that the 700mhz spectrum auction is over, wireless companies like Verizon and AT&T are going to start building high speed networks around that spectrum to provide instant high speed internet access to mobile devices.

AT&T won a bunch of licenses on the B block in major metro areas.
Verizon won a nationwide footprint (i believe it was the C block) on the 700mhz spectrum and is going to integrate their 4G LTE(Long Term Evolution) solution into it to provide theoretical downlink speeds up to 150mbits. The 4G solution is expected to go live in 2010. At that point you wouldn't need to worry about saturation and degraded service during peak times especially in congested areas or saturated sectors. That would allow them to get rid of the 5GB cap that they've put in place to try to preserve network quality for all users.

So I would guess that by the time these CoC's hit the market and mature a little bit, there should be a wireless internet infrastructure that would be able to handle the load that millions of these devices would place on it.
 
There's going to be mass changes about that in the future. Now that the 700mhz spectrum auction is over, wireless companies like Verizon and AT&T are going to start building high speed networks around that spectrum to provide instant high speed internet access to mobile devices.

AT&T won a bunch of licenses on the B block in major metro areas.
Verizon won a nationwide footprint (i believe it was the C block) on the 700mhz spectrum and is going to integrate their 4G LTE(Long Term Evolution) solution into it to provide theoretical downlink speeds up to 150mbits. The 4G solution is expected to go live in 2010. At that point you wouldn't need to worry about saturation and degraded service during peak times especially in congested areas or saturated sectors. That would allow them to get rid of the 5GB cap that they've put in place to try to preserve network quality for all users.

So I would guess that by the time these CoC's hit the market and mature a little bit, there should be a wireless internet infrastructure that would be able to handle the load that millions of these devices would place on it.

Also, CERN has built a 10gbps Trunk Network across the Globe to feed the LHC Project, that has been Taken on by other network Carries too and soon most will have 10gbps to even the smallest ISPs over the next few years.
 
Great posts! I've been coming to this site for years and it's helped me make a lot of major invesment strategies. Nowadays, this is one of the few places with unbiased information!! ...you guys here do actual research before you print a story AND you're not afraid to print a negative article (especially against Apple). Thanks again!
 
Great posts! I've been coming to this site for years and it's helped me make a lot of major invesment strategies. Nowadays, this is one of the few places with unbiased information!! ...you guys here do actual research before you print a story AND you're not afraid to print a negative article (especially against Apple). Thanks again!

heh :p
 
Great posts! I've been coming to this site for years and it's helped me make a lot of major invesment strategies. Nowadays, this is one of the few places with unbiased information!! ...you guys here do actual research before you print a story AND you're not afraid to print a negative article (especially against Apple). Thanks again!

If you were serious about investing you would be in Apple.

I made a fortune on it last year, made some of my friends money on it late last year, and just in the last three weeks I got a friend in at 118 and out at 152. Turned $20000 into about $26000 in three weeks, he owes me now. :) Per square foot the stores have most profit of any retailer in the country, all their products have huge profit margins, they have a customer base that continues to grow year over year, they have continues to show double digit growth in a time when computer sales are contracting, they are expanding into the smartphone market (something that alone has given RIMM a $60BN market cap), and they have great products. It is a great stock that will only continue to blow up. I would load up if it goes below $130 again.

All that said, all my short term cash is in commodities and currencies. I trade options, not stock, short term, and stock options in the short term scare the bejesus out of me right now. Now is a decent time to go long term in stock with the right beaten down companies. Short term, this stock market is too gnarly IMO we aren't out of the woods yet. I'll be back after summer is over.

And yes, I am long NVDA.
 
Back
Top