Petaflops

[geonuc]

I learned a new word today: petaflops. A new military computer processes over a quadrillion calculations per second, or one petaflops.

http://www.nytimes.com/2008/06/09/technology/09petaflops.html

[Wear Protection!]

incredible. nice article, thanks for sharing.

[hhEb09'1]

I learned a new word today: petaflop. A new military computer processes over a quadrillion calculations per second, or one petaflop.

http://www.nytimes.com/2008/06/09/technology/09petaflops.htmlFour more:The next thousandfold goal is the exaflop, which is a quintillion calculations per second, followed by the zettaflop, the yottaflop and the xeraflop.:)

A xeraflop is a trillion petaflops.

[Infinity Watcher]

Can I be mildly pedantic? (wait this is BAUT what am I asking for? ;) ) you do actually need the s on the end of flops: FLoating point Operations Per Second (http://en.wikipedia.org/wiki/Flops). Of course somebody who is actually trainedin computers is now going to come along and tell me why I'm wrong but I'm as curious as the next guy.

That doesn't diminish that this is a pretty cool machine, I do find it interesting that we seem to be shifting more towards increased parallel processing rather than increased processing speed of an individual chip, am I correct in making that observation?

[geonuc]

Can I be mildly pedantic? (wait this is BAUT what am I asking for? ;) ) you do actually need the s on the end of flops: FLoating point Operations Per Second (http://en.wikipedia.org/wiki/Flops). Of course somebody who is actually trainedin computers is now going to come along and tell me why I'm wrong but I'm as curious as the next guy.
Petaflops it is, then. :)

[hhEb09'1]

That doesn't diminish that this is a pretty cool machine, I do find it interesting that we seem to be shifting more towards increased parallel processing rather than increased processing speed of an individual chip, am I correct in making that observation?It's sorta both, but massive parallelism is getting more understood, and cooler both ways, so why not use what you got? Especially if it's cheap.

[HenrikOlsen]

Can I be mildly pedantic? (wait this is BAUT what am I asking for? ;) ) you do actually need the s on the end of flops: FLoating point Operations Per Second (http://en.wikipedia.org/wiki/Flops). Of course somebody who is actually trainedin computers is now going to come along and tell me why I'm wrong but I'm as curious as the next guy.
You're right, 1 flops would be correct, albeit rather slow.

[tlbs101]

Can I be mildly pedantic? (wait this is BAUT what am I asking for? ;) )

OK, here's another point-pedantic. The 'p' in petaflop(s) should be capitalized, especially if it is used as an abbreviated prefix. Lower-case 'p' is used for the suffix pico, meaning 1 trillionth. Suffix 'P' is used for Peta.

Imagine a computer that operated at 1 pflop. That's something like 1 calculation every 35 years.

It's 1 Pflop, not 1 pflop.
-------------------

I, too, think it is awesome that the world's fastest computer is 'right up the steet'. I also have a good friend that works for the lab in IT. I'll definitely be asking him if he has anything to do with this new machine.
.

[hhEb09'1]

OK, here's another point-pedantic. The 'p' in petaflop(s) should be capitalized, especially if it is used as an abbreviated prefix. Lower-case 'p' is used for the suffix pico, meaning 1 trillionth. Suffix 'P' is used for Peta. No, only if it is used in an abbreviated form. The term petaflops is not (should not be) capitalized, so far as I know. SI units (http://physics.nist.gov/cuu/Units/prefixes.html)

[Tobin Dax]

No, only if it is used in an abbreviated form. The term petaflops is not (should not be) capitalized, so far as I know. SI units (http://physics.nist.gov/cuu/Units/prefixes.html)
Just to help this point, gigabyte, megaparsec, etc., are not capitalized either, but the SI prefixes are capital letters.

[hhEb09'1]

Just to help this point, gigabyte, megaparsec, etc., are not capitalized either, but the SI prefixes are capital letters.Abbreviations, not prefixes. :)

[Tobin Dax]

Abbreviations, not prefixes. :)
Abbreviations/symbols for the prefixes is what I meant to say. :shifty: Oops. Thanks, Grapes.

[tlbs101]

No, only if it is used in an abbreviated form. The term petaflops is not (should not be) capitalized, so far as I know. SI units (http://physics.nist.gov/cuu/Units/prefixes.html)

I'm still a bit "old-school" on that subject, but I'll try and follow the latest-and-greatest from NIST from now, on. Thanks for pointing that out.

.

[hhEb09'1]

Y'all welcome, it's a pleasure here. :)

To tell you the truth, I've been through so many iterations, past and future, of this, it's hard to tell what is current. And I'm uncomfortable using NIST as a reference, in as much as it is an arm of a non-compliant. :)

[petaflops]

seems like as good a user name as anything else

[hhEb09'1]

Welcome to BAUT, peta! (always wanted to say that)

[grant hutchison]

I do like the concept of the kibibyte, for 1024 bytes, to be contrasted with a kilobyte, for 1000 bytes. It would sort out a real mess if people could be persuaded to use the IEC binary prefixes for powers of 1024, and SI prefixes for powers of 1000.
So we have kibi-, mebi-, gibi-, tebi-, pebi- and exbibytes.
Mebibytes would be a particularly appropriate measure for vapourware. :)

Grant Hutchison

[Nick Theodorakis]

I do like the concept of the kibibyte, for 1024 bytes, ...

Sounds like a fast-food version of a Lebanese ground meat dish.

Nick

[Larry Jacks]

As a historical comparison, the Cray 1 supercomputer (http://en.wikipedia.org/wiki/Cray-1) was built in the mid-1970s. It was a sensation at the time.

The system used twelve functional units, but had limited parallelism. It could fetch one instruction per clock cycle into the units, but operate on them in parallel and would retire two. Its theoretical performance was thus 160 MIPS (80 MHz × 2 instructions), although there were a few limitations that made floating point performance generally about 136 megaflops. However, by using vector instructions carefully and building useful chains, the system could peak at 250 megaflops.

If I got my exponents correct, the new computer is 4.1 million times faster than a Cray 1. Not too shabby!

[HenrikOlsen]

OK, here's another point-pedantic. The 'p' in petaflop(s) should be capitalized, especially if it is used as an abbreviated prefix. Lower-case 'p' is used for the suffix pico, meaning 1 trillionth. Suffix 'P' is used for Peta.

Imagine a computer that operated at 1 pflop. That's something like 1 calculation every 35 years.

It's 1 Pflop, not 1 pflop.
Again, it's Pflops/pflops, if you're talking per second.
If you're writing the suffix as peta, capitalization doesn't matter.

[Tobin Dax]

Again, it's Pflops/pflops, if you're talking per second.
If you're writing the suffix as peta, capitalization doesn't matter.
I've seen this twice now. "Suffix" may be appropriate for Henrik to use, but tbls101 should know better. A suffix comes after the root word, while a prefix comes before it.

[geonuc]

seems like as good a user name as anything else
Yer welcome. :)

[tlbs101]

I've seen this twice now. "Suffix" may be appropriate for Henrik to use, but tbls101 should know better. A suffix comes after the root word, while a prefix comes before it.

Arrrrrgh... prefix
I should know better.
I must have posted that before my 1st cup of coffee yesterday morning.
.