Posts by tpm 125 publicly visible posts • joined 1 Nov 2010
I was merely trying to describe what is. I actually admire both ARM and Linaro. You are reading tone that is not there.
To have made that more clear, had I not been rushed, I could have added this:
"This stands in stark contrast to the totalitarian hardware model of Intel, where you get what they make and like it, and Microsoft, where you get what they make and sometimes don't like it."
Now THAT would have been a snarky comment. No snarks intended at either ARM or Linaro. One makes money, one distributes money. Both are collectives and I think good ecosystem players.
See page 4 here:
http://www.stratus.com/Products/ftServerSystems/~/media/Stratus/Files/Library/Collateral/AutomatedUptimeLayerBrochure.pdf
I also contacted Stratus and triple-checked it and they confirmed that you only need one software license for the OS or hypervisor on a pair of machines.
Shooting seems a bit extreme. As the writer, I can tell you I have a bit of word poisoning after a year of this, so you might as well shoot me, too. I think perhaps some coffee would work better than bullets.
I mean, we don't shoot programmers when their code has a bug. Not yet, anyway.
Actually, I put the most recent data into the Top500 story here:
http://www.theregister.co.uk/2012/11/12/top_500_supercomputers_nov_2012/
Dell and TACC said Stampede would be around 9 petaflops, but as you can see in more current documents I cited at https://www.xsede.org/documents/234989/378230/XSEDE12_Stampede_and+MIC_Overview.pdf, the plan is to reach around 10 petaflops peak with a 6 megawatt power draw.
Well, there is another option. That there was no Sandy Bridge-EX because Ivy Bridge-EX was too close and that Ivy Bridge-EX and Maybe even Haswell-EX will go into what we would have called an Itanium socket.
I am not saying this is the case--Intel has not said. But putting Xeon E7 in an Itanium socket is just as plausible as the reverse.
K10 prices were not announced by Nvidia, but buy.com has one listed for $4,000. Which sounds about right based on the Fermi M2090 prices I have seen at server makers.
It must be tempting for Nvidia to think they can charge $10,000 for the K20. Three times the flops for three times the price is not a price/performance improvement. So I am thinking maybe $6,000 is a reasonable price if it can hit the performance levels El Reg is estimating.
We'll see. And a lot, of course, depends on Xeon Phi prices, which no doubt will be lower per flops than whatever Nvidia charges.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
