Upcoming Nvidia CPU (aka Denver)

Hows it going everyone? Has anyone heared of Nvidias' Project Denver? Supposedly Windows8 will support this cpu. I think it's cool that Intel might finally have some competition in the high end market again. I'm currently running a watercooled QX9650 @ 4ghz for gaming and it seems to be handling everything so far. The only upgrade I think I might be needing is an Nvidia Kepler 28nm when they roll out and then it's all about waiting for the "Denver" cpus to launch and get reviewed. I'm thinking by the time that happens Intels "Ivy Bridge" will be in its second generation, or the tock cycle of Intel and I'm wondering if Nvidia is going to blow their ship out of the water or will it just be somewhat competetive. What do you guys think?

Denver is an ARM instruction-set architecture

If its an ARM chip its likely going to roll out for tablets and other mobile devices first, and probably only as a dual core, since there really is no justification for a quad in small devices, and adding ARM support to windows is just a good idea in general as it allows for even larger market share.

It takes nVidia 6 months to a year to beat AMD's gpu's, and only then because they are much larger chips with higher power draw. In technology terms they cannot compete with AMD's graphic division which has less than half of nVidia's budget.

What makes you think this would suddenly change with a cpu? The best Nvidia would be able to do is match intel's gpu performance, ie not good enough by far.

From what I understand Nvidia has been working on this for a few years now and the information is just now coming out. Just like AMD they are going to the 28nm proccess so I'm guesssing they finally have the right tool for the job. Why is Bull dozer taking so long to roll out? Iv'e been hearing the name for years now. Nothing on the shelves.

"Project Denver"



"Project Denver" and Fermi.



Recoloured

No joke. I just didn't want to say it because it's my thread. Thanks by the way.

Don't know why you are thanking him -- and you are correct that it is your thread (which in the OP your supposition was Denver might blow Chipzilla's ship out of the water).

If that ain't drinking The Jen-Hsun Huang Kool-Aid, I don't know what is

The question you must ask: Why does Jen-Hsun Huang allow 'cooked' presentations and puffery to take away what may be a decent product?

DISCLAIMER: I personally think that nVIDIA is the worst company in the IT industry in terms of their "moral" and "ethical" conduct. I think I should point this out before posting... keep that in mind I am biased when it comes to nVIDIA

I think it will be decent for Tablets and Mobile Devices but I don't see it being good for desktops. Why? because nVIDIA make the worst chipsets this side of YUANG GIANG (never heard of YUANG GIANG?? that's because I made them up... nVIDIA is so horrible my imagination cannot even keep up with how bad they are at making chipsets).

I recently retired an AMD Athlon64 X2 3800+ Socket 939 nForce4-SLI rig because of nVIDIAs piss poor chipsets and chipset drivers.

I replaced it with a Pentium D 3.2GHz which outperforms the Athlon64 X2 under Windows Vista and 7 due to the horrible SATA/IDE drivers from nVIDIA.

So... that being said... how are nVIDIA going to implement driver cheats for their CPUs? Will they round out results in order to cut down on the processing times? Hehehe.

You should buy 3 gtx 480's and run them in tri-sli

Got it!

I laughed.

ElMoIsEviL - upgrade your phys-x card..
I run a GTS 240 (oem) as a phys-X card, single slot and one of the better performing single slot cards they got/ had at the time, still is..
main gfx card is GTX 560Ti.

^
GTX 460 for phys-x...!
damn.

If its anything like their mobo chipsets, I wont buy it.

I dont agree their GPU are bad but I do agree their Mobo design is really bad.
To me they have a very strong postion in the GPU share.Why not go for it?
We need more competition!

What nVidia is doing: Combining ARM with a graphics engine.

To promote their next generation they 'canned' some benchmarks (in a measurement that actually made no sense) and compared it to a T7200.

They used the current compiler for their chip, claiming it was faster than the T7200. The problem is, they used a compiler from 2004-2005 for the T7200.

Enterprising enthusiasts ran the T7200 under the current compiler and, "Surprise! Surprise!" -- The T7200 was 50% faster than the nVidia chip.

That's pretty much the definition of ...

^
not necessarily..
just matter of pre-release public opinions.

that's why I keep saying on every thread I post on:
I want to see the performance benches..

Project Denver is an ARM architecture CPU and I like the ARM architecture. It is superior to current X86 based CPUs in some respects, but of course all our current PC programs are for x86 32/64 bit CPUs ^_^. ARM may or may not become the dominant architecture in the long run, but I can't see it happening anytime soon unless both Intel and AMD drop the ball big time.

It's a type of RISC architecture. It stands for Advanced RISC Machine or something. It was originally used for Acorn desktop computers, which never caught on, and then was spun off to basically license out the ARM architecture. Look at the wikipedia page if you want to know more.

http://en.wikipedia.org/wiki/ARM_architecture

ARM is definitely more flexible than x86 (google RISC vs CISC). However, scalability is a factor, which i think ARM also has in its favor. Problem is, it will take probably to around the year 2015 for ARM to really compete with x86 on a full market basis, if ever.

This.

Intel is expecting to not only show off Ivy Bridge in May but also a new Atom. My guess would be a 22nm refresh or the smart phone version.

Intels 22nm was shown pushing 3.8GHz on less than a volt (0.95 if I remember) which means that at the same voltage as a ARM CPU, it can be clocked higher or they can add more cores.

AMD has the same advantage since they have a cross license with Intel so they can see Intels process tech.

Problem is, with power efficiency in mind "Any thing x86 can do, ARM can do better, anything x86 can do, ARM can do too". ARM SoC's at 22nm could probably draw half of what they do today...or be 2-3x faster. x86 will probably not get into the SoC market for another 3-5 years, and ARM probably wont hit full power devices for another 3-5 years. Thats where the real competition starts. But ARM is the one i think wins. Reason isnt an argument of RISC vs CISC, its an argument of what consumers want. Nobody has a Mac anymore, they have an iPhone, iPad, or iMac. Linux is good an all, but how many people use it as their regular OS? Very few...like between 0 and 1% id imagine. But how many people use an Android smarthphone (Android being based off and coded from Linux), yeh thats more than iOS and Blackberry combined last time i checked. How many kids ask their parents for a new desktop? How many ask for a new phone, laptop, or tablet? Fact is, things are migrating to the mobile market. Mobile market is where battery life counts. Battery life is dependent on primarily power consumption. Power consumption is based primarily on how much power CPU, GPU and whatever else that would be found on a SoC that uses power. For that reason, and the fact ARM is more efficient than x86, im going to pick ARM in this fight.

Lolz... I love how everyone says Nvidia is trying to compete with Intel when it was the other way around. Intel has been trying to step on Nvidias shoes with the built in GPUs for some time now. It was only a matter of time before both companies started to make a combination CPU and GPU combo.

What I can gather is this:

1) Nvidia will not be able to compete with Intel/AMD in the server market at all - ever. They will most likely make a cheap low-power low-ability netbook type APU. It will probably sell well and create a new genre of chips that will be mainstream.

2) Intel will continue their current scheme of tick-tock rape everything and not change anything in their plans.

3) Nvidia will make most of their profits from selling high brand GPUs.

4) AMD will still be screwed in the long run unless bulldozer can surprise us all - which I doubt it will.