Replies to post #98648 on Intel Corp (INTC)

[mmoy]

In general, where you can compete depends on the workload.

There are workloads that run well on GPUs from nVidia right now. But you have to do specialized programming to make good use of those resources. You might want to grab K&R C, and learn how to program along with the difference between integer and floating point operations. This is a first step in gaining some understanding about the issue.

[herb will]

"NVidia have to build a processor that competed in ALL of Intel's markets"

I try to stay out of these sort of speculations because it is totally out of my bailiwick but I find the following comments by the ARMH CEO informative.

Warren East
Q410 CC
“They’re extending their Tegra family around Cortex-A15, but also they’ve become an architecture licensee for new products smoothing out in the computing spectrum. So taking ARM right across the computing spectrum in future, that’s their plans. “
“Well, really we – it’s a bit early to start for us to start speculating too much on exactly where ARM is going to get to in the computing space. I mean we’re saying there is a potential there and what we are highlighting on the chart was actually the potential for the Cortex-A processor. “

Q310 CC
“So we’ve said that all we can say about that agreement on its own. More generally about what the ARM partnership is doing in this space. Right now, we’re seeing people experiment with the chips. Chips developed to-date are targeted at high-end smartphones, they’re not ideal for server applications. And so we’re seeing semiconductor companies experimenting by building different system chips around our microprocessors cores to target that space.
The one agreement announced about the architecture licensee is just one example of that. We’re not really, therefore, expecting any material impact as far as royalty is concerned, for several years, hence I think this is a 2013 by the time we see the server chips hitting the market. And then we’re going to have design-in ramps and so on. So it’s going to be really something, before there’s a material movement it’s going to be in the sort of five years and beyond period.”
http://seekingalpha.com/article/249992-arm-holdings-ceo-discusses-q4-2010-results-earnings-call-transcript
http://seekingalpha.com/article/232356-arm-holdings-ceo-discusses-q3-2010-results-earnings-call-transcript

[chipguy]

Maybe but the tide of history in computing is going against it. HPC
has been moving away from special purpose, low volume silicon
for quite a while. Some of the recent top500 machines have padded
their flop numbers by complementing their Xeons with GPUs but
the reality is real HPC work is a lot harder than simply cranking out
Linpack numbers. Folks who customize apps for these machines
are finding out after a lot of hard work the best results are usually
obtained by ignoring the GPUs and running apps just on the Xeons.
Given the politics of funding /dick size contest nature of the top500
list the use of GPUs in top end supers won't end any time soon but
the reality that it is hard to gain much or any extra performance by
adding Nvidia GPUs for a lot of HPC problem classes is becoming
more widely known. IMO Intel has a better strategy for HPC with
its Larrabee platform of x86 based CMP chips optimized for vector
processing but with a general purpose, coherent cache hierarchy.
Again here Intel has the expertise to pull it off. It is likely far, far
easier to adapt HPC apps to that sort of platform than to a mish
mash of Xeons and GPUs.