Headline FeedsTechNewsWorld Talkback
|
|
See Full Story
For the last decade, the PC market has kind of sucked. Against the massive growth of the '90s, and with the exception of Apple, which didn't do well in that decade, the PC market has been a poor reflection of the excitement that once surrounded it. Part of the problem was the focus on computation and the lack of focus on things that make computing exciting. In this respect, Apple was the exception, and that helps explain why users tend to be more excited about Apple products. They have, until now, virtually stood alone -- but that is about to change.
I like your article overall but I'm going to disagree with your optimism concerning Nvidia.
The concepts of PhysX, CUDA and Tesla are all great but about 10 years too late. With the pending release of DX 11/OpenCL most of what the NV versions accomplish will be rendered obsolete virtually overnight. At least as far as it requires you to have an Nvidia product.
In order for Ion to take off, they are going to have to overcome the marketing practices of Intel so they can compete against Atom. AMD will be releasing the Neo. AMD also has Fusion and given that Nvidia has no x86 path they will be unable to even enter that market.
That basically leaves the high-performance/enthusiast segment open to Nvidia where they are struggling desperately to get their act together on that front.
When you add in their ongoing issues with manufacturing, recalls and history of manhandling customers and partners their future seems bleak at best.
The concepts of PhysX, CUDA and Tesla are all great but about 10 years too late. With the pending release of DX 11/OpenCL most of what the NV versions accomplish will be rendered obsolete virtually overnight. At least as far as it requires you to have an Nvidia product.
In order for Ion to take off, they are going to have to overcome the marketing practices of Intel so they can compete against Atom. AMD will be releasing the Neo. AMD also has Fusion and given that Nvidia has no x86 path they will be unable to even enter that market.
That basically leaves the high-performance/enthusiast segment open to Nvidia where they are struggling desperately to get their act together on that front.
When you add in their ongoing issues with manufacturing, recalls and history of manhandling customers and partners their future seems bleak at best.
Each of the Major Vendors Has Critical Limitations
Posted by: RobEnderle 2009-05-18 17:05:11 In reply to: ir0x0r
With NVIDIA you are correct they lack the x86 base they need to complete a system and both Intel and AMD are moving to an integrated model which could lock them out. On the other hand while DX 11, and OpenCL are coming they aren't here yet and by moving on Cuda NVIDIA is capturing a lot of the initial interest and it is generally easier to keep developers than to capture them in the first place. On Physics you have the Intel/Hazard stuff and PhysX, I have yet to see a compelling game that uses either so can't call a winner here yet and AMD might be better off licensing a technology that Intel didn't own given Intel is killing cross licenses right and left with both NVIDIA and AMD at the moment. What NVIDIA does have is breadth from servers all the way down to hand held computers with automotive in the middle suggesting they are the one to beat on a graphics pure play for the future.
AMD has the x86 part that NVIDIA misses and is creating a very interesting Fusion Platform, they lack the product reach that Intel and NVIDIA have but they make up for that with a tighter platform focus. The problem is the solution spans from handhelds to servers and AMD no longer has the handheld space. So they are stronger on PCs and potentially (talking graphics) stronger on servers, but they don't have the reach needed to complete the solution and drive the market.
Intel owns x86 but that actually hinders them in graphics as this is like a religious argument for them and I worry that their impressive Digital Home effort will get stomped for being on the wrong side of the internal Intel power curve. Clearly Intel has the breadth and is better funded than either NVIDIA or AMD but their excessive focus on x86 has kept them from seeing this opportunity emerge until recently and their heavy focus on litigation activities suggests they know they are out of step but are uncertain how to correct the problem.
So, you are correct, that NVIDIA has issues but so do the others. I just think NVIDIA has the more complete vision and were it not for the x86 shortcoming (they do have ARM) I'd give them the edge. As it is, I think it is anyone's game for now. NVIDIA has the better vision, AMD the most aggressive merged technology strategy, and Intel the strongest positional power and most resources. Whichever eliminates their shortcomings first likely wins this race.
AMD has the x86 part that NVIDIA misses and is creating a very interesting Fusion Platform, they lack the product reach that Intel and NVIDIA have but they make up for that with a tighter platform focus. The problem is the solution spans from handhelds to servers and AMD no longer has the handheld space. So they are stronger on PCs and potentially (talking graphics) stronger on servers, but they don't have the reach needed to complete the solution and drive the market.
Intel owns x86 but that actually hinders them in graphics as this is like a religious argument for them and I worry that their impressive Digital Home effort will get stomped for being on the wrong side of the internal Intel power curve. Clearly Intel has the breadth and is better funded than either NVIDIA or AMD but their excessive focus on x86 has kept them from seeing this opportunity emerge until recently and their heavy focus on litigation activities suggests they know they are out of step but are uncertain how to correct the problem.
So, you are correct, that NVIDIA has issues but so do the others. I just think NVIDIA has the more complete vision and were it not for the x86 shortcoming (they do have ARM) I'd give them the edge. As it is, I think it is anyone's game for now. NVIDIA has the better vision, AMD the most aggressive merged technology strategy, and Intel the strongest positional power and most resources. Whichever eliminates their shortcomings first likely wins this race.
