IBM Brings Supercomputing Muscle to U.S. Lab

All Things Digital ^ | February 8, 2011 at 7:06 AM PT | Arik Hesseldahl

[Paul Ross]

A better question would be, "What are they doing with it?"

Here were some suggestions for the possile "what"

can you do with 10 or 20 petaflops? Meteorologists could predict local weather down to the 100-meter range with a 20-petaflop system. And running a simulation of how a beating human heart reacts to new medicine, which takes two years of computing time today, will get done in two days on a 10-petaflop system.

Somehow I think we will be doing less productive things.

[Marine_Uncle]

I am sure the research scientist are lined up to get time on these future machines.

[jdsteel]

Can you screen all cell and telephone communications for key words and spit out the non-PC offenders names? ; )

[HeartlandOfAmerica]

Could handle an awful lot of pron on a machine like that! :)

[SunkenCiv]

Thanks Ernest!

ps/3 supercomputer site:freerepublic.com

Google

[bmwcyle]

ping

[grey_whiskers]

I second the motion.

Peak teraflop ratings on LINPACK or SCALAPACK or whatever or nice, but if the bulk of your code relies on algorithms that aren't parallelizable, or your data partitioning isn't friendly, you get bit on the ass by Amdahl's Law.

Haven't heard much about robust numerical methods which map to these architectures -- but of course theoretical chemists would just *love* larger basis sets, even with the returns limited by the non-linear scaling.

Cheers!

[NormsRevenge]

Maybe they can set up some solar energy fields just outside of Livermore to power the new computers.

Otherwise they might need a coal powered generating station...

—

Heck yeah,, they have acres of wind turbines nearby at Altamont Pass.. I bet birds would be glad to see them go away..