sure, your solution may win if you don't care about bandwidth between separate nodes/OS images.
if you do, however, Cray's backplane is much faster than anything you'll wire up yourself, and is easier/takes up less space then a rack full of infiniband-ed 1U's.
As others have said, most of the MP supercomputers (IBM, Cray, et. al) fit this definition of 'turnkey cluster.' Yes, you do use MPI and you use a job scheduler to handle the details of placing jobs larger than one node on multiple 'compute' nodes.
Re:You want me to pay for that?
on
The Universal Card
·
· Score: 2, Interesting
yeah, just what i need. one single device that can decide to call 'home' with all of my purchasing habits on it.
don't bother compiling mozilla/firebird/openoffice et. al. download the official binaries from their respective sites (and emerge libcompat if needed) you have almost nothing to gain by compiling those from source.
auburn university is using a cisco vpn solution to secure the node-to-access-point communications. the vpn client is available for windows, macos, and linux.
i do a similar thing, but we use the picture of a southern-fried chicken head that somebody supposedly found in a mcdonald's chicken mcnugget box. people usually don't let the chicken head stay.
also, we have select biographies (with full pictures) online, and there some "honorary" members of our organization. anna kournikova and natalie portman are my favorites there.
i'm sorry but they do. and anyway, who designs the next xVM that talks to the new hardware? i am defending this position because i'm an embedded developer and programmers:
1) sure as hell do need to know what's going on at least in the general case, and
2) the market for commercially developed software will change and there will still be programming jobs in the US, even if the "brainless" work is offshore.
and on what software does the "point-and-drool" tools run? and who writes that?
anyone who has taken an automata and computability course at college level can tell you that such software is non-trivial to write/maintain.
you paint a nice, dismal view of the industry in 15 years, and as you say SOME jobs may be moved offshore. what do i care if some "point-and-drool" master coder is out of a job because of it?
by 2015 computers will be fast enough that point-and-drool paint-by-numbers tools will be available to rapidly and idiotically autogenerate most of the code you write today with no discernable performance loss
no, but let's face it... life in space is tough. europa and any other lump of rock out there that may be harboring life are all subject to the same harsh conditions. it's "good science" to try not to contaminate one's observations by one's own mistakes.
from KJV... i knew if i quoted the n.e.t. or n.i.v. bible, then i'd be discredited on that basis...
EXODUS 21:
22 If men strive, and hurt a woman with child, so that her fruit depart from her, and yet no mischief follow: he shall be surely punished, according as the woman's husband will lay upon him; and he shall pay as the judges determine.
23-25 And if any mischief follow, then thou shalt give life for life, Eye for eye, tooth for tooth, hand for hand, foot for foot, Burning for burning, wound for wound, stripe for stripe.
what does "and yet no mischief follow" mean? the foot note here in many bible translations say that the original (hebrew) seems to mean that the child is born and neither it nor the woman dies.
granted that this is quite a battle for evangelical Christians versus (almost) everyone else, but you make it sound like the bible says that if you kill the unborn you should just pay for it and it's cool.
what major should I choose? It seems that many companies are looking for computer scientists, but would they be desperate enough to accept computer
engineers? What is the difference anyway?
being a senior in CS and a COOP with a mostly EE/CE company, the job you get doesn't really hinge as much on what your degree is in as much as who you know and what you can do. i am interested in algorithm analysis and applied discrete math and such, but the EE's at my company couldn't care less about such things!
at my company, we build embedded networking appliances, which entails designing the hardware and writing the embedded software to drive it. they used to have 5 EE/CE's to about 1 CS person. the EE/CE's would build prototypes and play with low-level bit flipping and they'd have the CS guy write all of the software. nowadays, however, those CS people have gotten better jobs elsewhere and have not been replaced, and we have a lot of EE's writing ugly code for their own devices!
the divide between CE/CS is especially evident in my situation, because some areas/projects here suffer for lack of good CS people. both are quite obviously needed for my company to turn out a quality product line. it all depends on what you like to do!
Slashdot Mirror
actually, there were 5:
http://en.wikipedia.org/wiki/Istari
http://en.wikipedia.org/wiki/Blue_Wizards
sure, your solution may win if you don't care about bandwidth between separate nodes/OS images.
if you do, however, Cray's backplane is much faster than anything you'll wire up yourself, and is easier/takes up less space then a rack full of infiniband-ed 1U's.
As others have said, most of the MP supercomputers (IBM, Cray, et. al) fit this definition of 'turnkey cluster.' Yes, you do use MPI and you use a job scheduler to handle the details of placing jobs larger than one node on multiple 'compute' nodes.
yeah, just what i need. one single device that can decide to call 'home' with all of my purchasing habits on it.
don't bother compiling mozilla/firebird/openoffice et. al.
download the official binaries from their respective sites (and emerge libcompat if needed)
you have almost nothing to gain by compiling those from source.
auburn university is using a cisco vpn solution to secure the node-to-access-point communications. the vpn client is available for windows, macos, and linux.
now i don't have to feel guilty about buying a pentium-m notebook as opposed to waiting for transmeta to finally deliver the astro!
i do a similar thing, but we use the picture of a southern-fried chicken head that somebody supposedly found in a mcdonald's chicken mcnugget box. people usually don't let the chicken head stay. also, we have select biographies (with full pictures) online, and there some "honorary" members of our organization. anna kournikova and natalie portman are my favorites there.
what's an lds? a mormon?
Is this what happens when you give one company a license to print money?
is this a rhetorical question?
You know why? Java runs on the --- JVM!!
the writers of the JVM DID need to know about their respective hardware platforms and
(and this is my point...)
the java developer NEEDS to know how the VM does things as well. it's a virtual machine... virtual hardware!
i'm sorry but they do. and anyway, who designs the next xVM that talks to the new hardware? i am defending this position because i'm an embedded developer and programmers:
1) sure as hell do need to know what's going on at least in the general case, and
2) the market for commercially developed software will change and there will still be programming jobs in the US, even if the "brainless" work is offshore.
and on what software does the "point-and-drool" tools run? and who writes that?
anyone who has taken an automata and computability course at college level can tell you that such software is non-trivial to write/maintain.
you paint a nice, dismal view of the industry in 15 years, and as you say SOME jobs may be moved offshore. what do i care if some "point-and-drool" master coder is out of a job because of it?
by 2015 computers will be fast enough that point-and-drool paint-by-numbers tools will be available to rapidly and idiotically autogenerate most of the code you write today with no discernable performance loss
so... who writes the "point-and-drool" tools?
no, but let's face it... life in space is tough. europa and any other lump of rock out there that may be harboring life are all subject to the same harsh conditions. it's "good science" to try not to contaminate one's observations by one's own mistakes.
well really i take it to mean that the child is born prematurely but not dead. how can a woman deliver a dead baby and "no mischief follow?"
like i said, it's not too clear and the point of conflict for a virtual war over biblical interpretation.
my only point was that the bible does not (explicitly or implicitly) say:
dead baby, live woman: $100
dead baby, dead woman: you are in big trouble!
from KJV... i knew if i quoted the n.e.t. or n.i.v. bible, then i'd be discredited on that basis...
EXODUS 21:
22 If men strive, and hurt a woman with child, so that her fruit depart from her, and yet no mischief follow: he shall be surely punished, according as the woman's husband will lay upon him; and he shall pay as the judges determine.
23-25 And if any mischief follow, then thou shalt give life for life, Eye for eye, tooth for tooth, hand for hand, foot for foot, Burning for burning, wound for wound, stripe for stripe.
what does "and yet no mischief follow" mean? the foot note here in many bible translations say that the original (hebrew) seems to mean that the child is born and neither it nor the woman dies.
granted that this is quite a battle for evangelical Christians versus (almost) everyone else, but you make it sound like the bible says that if you kill the unborn you should just pay for it and it's cool.
being a senior in CS and a COOP with a mostly EE/CE company, the job you get doesn't really hinge as much on what your degree is in as much as who you know and what you can do. i am interested in algorithm analysis and applied discrete math and such, but the EE's at my company couldn't care less about such things!
at my company, we build embedded networking appliances, which entails designing the hardware and writing the embedded software to drive it. they used to have 5 EE/CE's to about 1 CS person. the EE/CE's would build prototypes and play with low-level bit flipping and they'd have the CS guy write all of the software. nowadays, however, those CS people have gotten better jobs elsewhere and have not been replaced, and we have a lot of EE's writing ugly code for their own devices!
the divide between CE/CS is especially evident in my situation, because some areas/projects here suffer for lack of good CS people. both are quite obviously needed for my company to turn out a quality product line. it all depends on what you like to do!
a little bit of sand in a plastic bag or other junkmail envelope does the trick.