Nvidia "Project Denver" ARM for Desktop

Of course it means nothing until it really exists :) I am looking forward to seeing some specs on it though....
 
I do not think this will mean anything serious for gaming. X86 is full of power using features for a reason and you can't just expect serious game developers to give up using these features which will cause a massive drop in performance in their games.

Where the danger potential (Especially for Intel) business impact is if this causes a major change in the set top box market. The Boxee box ditched Tegra 2 because that core failed to handle many needed tasks. If you changed that .5W core into a 5W system with additional features and of course ways to render content you will have a serious competitor to Ontario and Atom.

However that will require that the price for these chips end up very low. And that they actually live up to expectations.
 
"VentureBeat" posted this on Youtube

http://www.youtube.com/watch?v=xKc_XGuvNIk

Microsoft is serious about this. And Nvidia had an impressive demo using just Tegra II

If I were Intel right now I would be panicking. This could cost Intel many orders for their X86 based systems for set top boxes.
 
Oak Trail will come to the front to address certain scenarios, and more follow up down the road...

The desktop and server parts are for now fantastic, more to follow...

Do not underestimate the collective intelligence/will/capability of Intel
 
Yeah, I don't think this is a stretch for nvidia. Instead of just using a licensed ARM core like in the Tegra series, it will be customizing a licensed ARM core. Dozens of companies do the same.

I just wonder if there are specific features that MS is suggesting to partners? I imagine common things like media decoding and an easy to interface audio processor (please not AC'97 :p) to have a minimum level of functionality. Or maybe even something like suggested/provided USB and SATA interfaces so Windows 8/ARM doesn't have a big mess of odd hardware to support. Remember that MS recently became an ARM licensee and can design and sell parts too, perfect for development.
 
More competition is most always good for the customer. Besides Intel has been encroaching in Nvidia territory for a while, and Nvidia was going to strike back.

The whole tech worlds seems to be a mess at the moment, going to be an interesting 5+years plus to come before the dust settles.
 
Lol I was just gonna post this story also.

Its looking good for competition again for sure!
 
More competition is most always good for the customer. Besides Intel has been encroaching in Nvidia territory for a while, and Nvidia was going to strike back.

The whole tech worlds seems to be a mess at the moment, going to be an interesting 5+years plus to come before the dust settles.



Unless you mean consumers of set top box systems I highly doubt this will provide more "competition" for most of us.

Again the main issue is for games and projects that are not just internet or media oriented. They NEED the extra X86 features so you wont suddenly see BC2 or CoD just ported.

Also I HIGHLY doubt they will spring for this over the 9W fusion core. For just 4W more you get X64 and the complete PC experience.
 
Unless you mean consumers of set top box systems I highly doubt this will provide more "competition" for most of us.

Again the main issue is for games and projects that are not just internet or media oriented. They NEED the extra X86 features so you wont suddenly see BC2 or CoD just ported.

Also I HIGHLY doubt they will spring for this over the 9W fusion core. For just 4W more you get X64 and the complete PC experience.

I think you are missing the significance. This IS a full PC experience now, with Windows 7/8 running on the ARM architecture.
 
I am excited about this. With Intel pushing their integrated graphics, and AMD producing CPUs/GPUs, I like the idea that Nvidia will be a CPU/GPU company. The big three.
 
No we won't get BC2 or CoD right away. What we will get is an amazing amount of competition.
 
I've been saying for a while that this day was coming. This is why the big push with .NET. Portability between Phone, XBOX, and PC no matter the architecture.
 
Unless you mean consumers of set top box systems I highly doubt this will provide more "competition" for most of us.

Again the main issue is for games and projects that are not just internet or media oriented. They NEED the extra X86 features so you wont suddenly see BC2 or CoD just ported.

Also I HIGHLY doubt they will spring for this over the 9W fusion core. For just 4W more you get X64 and the complete PC experience.

I think you overestimate how significant gamers are as far as computers... most computers live out their life in a cubicle, where a lower power machine is always welcome... and lower cost even more welcome...

99% of people running office do not care about having a quad core i7, just how quickly their "email" aka outlook opens ;)

and thats where I think this could really, really hurt intel
 
I, for one, am hoping that this catches on. X86 is where the performance is, but the architecture itself is old, outdated and just downright crap. As far as I'm concerned, Intel and AMD are pouring money into a sinking ship all in the name of "legacy support."
 
Unless you mean consumers of set top box systems I highly doubt this will provide more "competition" for most of us.

Again the main issue is for games and projects that are not just internet or media oriented. They NEED the extra X86 features so you wont suddenly see BC2 or CoD just ported.

Also I HIGHLY doubt they will spring for this over the 9W fusion core. For just 4W more you get X64 and the complete PC experience.

What is with your fetish with set top box systems? Most set top boxes aren't X86 systems now, why should this have any impact on that whatsoever?
 
99% of people running office do not care about having a quad core i7, just how quickly their "email" aka outlook opens ;)


for that, i'm thinking a modern SSD would help more than a new processor.

most of the boxes at work have low end C2D chips, but they seem infinitely slower due to the 5200rpm HDDs (and 9000 layers of security). drives me nuts now that i'm used to SSDs at home.
 
for that, i'm thinking a modern SSD would help more than a new processor.

most of the boxes at work have low end C2D chips, but they seem infinitely slower due to the 5200rpm HDDs (and 9000 layers of security). drives me nuts now that i'm used to SSDs at home.

This has been my contention as well. I/O is more of a bottleneck than processing power. Switching from a mechanical drive to a decent SSD is like night and day. Not to bring Apple into this, but playing around with the new Air adds validity to this point; Its snappy considering it has a dated Core 2 Duo CULV part in it.

It will be interesting what Nvidia can do with ARM in the desktop arena, however.
 
Nvidia will be using their Tesla Gpus as the heart of the system and use the Arm chips as the traffic cop to make sure all the data is going where its suppose to go and to do any cpu optimized calculations
 
I agree with those that think this will have little impact on the desktop space or home users. Regardless of whether Windows is ported to ARM, it is the vast Windows x86 software library that makes Windows what it is. Just because you port Windows to ARM, doesn't mean that existing Windows programs will run. Look at how long it took Windows x64 to catch on, and even that never would have caught on if it didn't have an x86 32-bit subsystem to run 32-bit apps.

With the ARM version of Windows there will be no x86 compatibility whatsoever. The only space I can see this getting a foothold *IF* their hardware is compelling enough, is the custom application server market.
 
1. Since there will be 100's millions of upcoming Cortex-derived tablets. Microsoft is unable to watch Apple iOS and Google Android devices gaining from single core to dual-core to quad-core with ever-increasing complexity and fully feature software ecosystem running full-blown applications without Microsoft's software presence.

2. Because mobile phone business is financed by 2-years contract, every 2-years someone is going to demand new thing, there will be continuous incentive and supply to meet this condition.

3. Microsoft decides it is going to participate by supporting the Cortex-ecosystem. Because mobile devices are always in renewing stage (2-years contract, at the end, many users want a new subsidized experience), the technology to supply always get advancement, meaning mobile-devices will get firmer and firmer ground. The software application sets in further and further.

4. One day, the people having large collection of those mobile software application/services will want the same experience on desktop.

It does not matter much that they are not the top performing parts. Look at the car industry. Some people demand Ferrari and Mercedes, others are perfectly happy to have an entry-level cheap basic safe car to drive around the city. Even more metropolitan people ride on public transports.
 
1. Since there will be 100's millions of upcoming Cortex-derived tablets. Microsoft is unable to watch Apple iOS and Google Android devices gaining from single core to dual-core to quad-core with ever-increasing complexity and fully feature software ecosystem running full-blown applications without Microsoft's software presence.

2. Because mobile phone business is financed by 2-years contract, every 2-years someone is going to demand new thing, there will be continuous incentive and supply to meet this condition.

3. Microsoft decides it is going to participate by supporting the Cortex-ecosystem. Because mobile devices are always in renewing stage (2-years contract, at the end, many users want a new subsidized experience), the technology to supply always get advancement, meaning mobile-devices will get firmer and firmer ground. The software application sets in further and further.

4. One day, the people having large collection of those mobile software application/services will want the same experience on desktop.

It does not matter much that they are not the top performing parts. Look at the car industry. Some people demand Ferrari and Mercedes, others are perfectly happy to have an entry-level cheap basic safe car to drive around the city. Even more metropolitan people ride on public transports.

a post after my own heart
 
Just graduated in Computer Eng. Looking at NVIDIA hard now. They are in a pretty good place afaik. Also need to look at what MSFT is going to be doing with this in the mobile space. Can anyone say legitimate portable gaming?
 
I think Intel definitely has a challenge on their hands here. We will have to see how the new Atom in 2013 performs, but by then it might just be a bit late...

However it's not like Intel hasn't come out from lagging behind the competition before.....
Personally, I am excited to watch all of this unfold. I think there is going to be a lot of cool stuff coming out in this space soon.
 
ARM is a splendid architecture; great performance per watt and performance per clock...the question is will developers take the jump?
 
Just graduated in Computer Eng. Looking at NVIDIA hard now. They are in a pretty good place afaik. Also need to look at what MSFT is going to be doing with this in the mobile space. Can anyone say legitimate portable gaming?

'Looking at Nvidia' for... ? Employment? Purchasing Stock? Buying your next PC?

I'm kind of curious myself as to how this is going to affect Nvidia's stock price / bottom line and, in the long term, affect their desktop GPU architecture. When combined with claims like this (taken with a large grain of salt), I wonder if there is going to be a mini GPU/PC Gaming renaissance in the next few years, this time focused on mainstream and mid-level performance, rather than the high end that's characterized PC gaming thus far.
 
Wait, wait. So NVIDIA using ARM cores, which have never been designed for high power or performance, will suddenly release a general purpose CPU, after having NEVER DESIGNED A HIGH PERFORMANCE CPU EVER, that will threaten veterans like Intel and AMD?

Yeah, that makes sense. As much sense as expecting a $250 FX chip to outdo a $999 Gulftown. Yup.
 
Wait, wait. So NVIDIA using ARM cores, which have never been designed for high power or performance, will suddenly release a general purpose CPU, after having NEVER DESIGNED A HIGH PERFORMANCE CPU EVER, that will threaten veterans like Intel and AMD?

I'm going to say you may be imagining things that haven't actually been announced.
 
Last edited:
This has been my contention as well. I/O is more of a bottleneck than processing power. Switching from a mechanical drive to a decent SSD is like night and day. Not to bring Apple into this, but playing around with the new Air adds validity to this point; Its snappy considering it has a dated Core 2 Duo CULV part in it.

It will be interesting what Nvidia can do with ARM in the desktop arena, however.

My opinion based upon everybody that has tried this has been this.
Video Chip Manufacturers can't make Desktop Processing Chips
Desktop Processing Chip manufacturers can not make Video Card Chip Makers
Low Powered Chip Manufacturers can't make high powered desktop processing chips.
High Powered Desktop Processing chip Manufacturers can't make Low Powered Chips.
Niche Server Chip Manufacturers can't make general purpose server chips.
General Purpose server makers can't make Niche ServerChip Manufacturesr.

Almost every single one of those combinations has been proven by NVIDIA/ATI/SPARC/Intel/ARM. So I will wait with very baited breath when a Math/Video Processing claims to make a desktop chip. I personally would put money on that project faring similar results to Larrabee. Perhaps Nvidia can do a bit better and actually get something out but neither desktop / general purpose computing nor power /efficiency are 3 things that Nvidia has ever produced aside from the semi-power efficient tegra2.

Imho given how many massive failures and how few actual success stories there have been companies shouldn't try crossing disciplines as much as they insist on doing. Instead focus R&D on improving their strengths.

On the flip side, for several years now, I've always wanted to see what an ARM powered desktop would be like. E.g. how does the instruction set compare to our beloved x86 instruction set. Could new additions to ARM be created to better solve more common problems that can't be done on the x86 platform because there are 20+ OSes that refuse to rewrite from the scratch.

So I guess we'll see what happens. 1 year down 2-3 more to go.
 
Wait, wait. So NVIDIA using ARM cores, which have never been designed for high power or performance, will suddenly release a general purpose CPU, after having NEVER DESIGNED A HIGH PERFORMANCE CPU EVER, that will threaten veterans like Intel and AMD?

Yeah, that makes sense. As much sense as expecting a $250 FX chip to outdo a $999 Gulftown. Yup.

Possibly true, but we should give them some room for possible success. If they are able to produce something that fits a niche, unlike Larrabee which failed to do so, then it might not be what they expected but it went somewhere.

And FYI. AMD is simply using the ARM Instruction set. They are developing their own microprocessor.
http://www.xbitlabs.com/news/cpu/di...Not_Interfere_with_ARM_s_Own_64_Bit_Tech.html
 
Back
Top