I am not a biochemist either, but I believe that certain types of molecular dynamics simulations and protein folding simulations can be parallized fairly efficently and do not need huge data sets.
Its not a factor of how many operations that need to be performed, its a matter of how many can be performed INDEPENDANTLY of the other processing nodes.
Look at IBM's Blue Gene supercomputer which is using relativley small capacity PIM (processor-in-memory) chips to do large protien folding calculations.
BTW, IANAB (I am not a biologist), so I can't say exactly what sort of simulations they are doing...
Who cares how many servers are deployed? It is irrelevant to EVERYONE. Your decision to use an OS should be completely independent of what other people are doing, right?
Idealisticly, yes. Realisticly, NO.
If I use an OS (or any product) which is commonly used then I will have a better chance of knowing that it will be supported by more vendors, more developers will make programs for it, and there will be a support community for it in the future.
Knowing the usage statistics for a platoform are important in that they allow you to know "if I go with product X, will I be stranded and forced to migrate in 3 years? or will there be a growing community developing and advancing the product?"
The quote that "the idea is to get rid of the keyboard" did _not_ come from Apple. It came from a source "familar with Apple's plans" which could mean just about nothing.
I doubt that apple would scrap keyboards on all of their powerbooks, at most they may offer one or two models without and add handwriting recognition to the rest.
It looks like they used Solaris x86, which is not as finely tuned as Solaris/SPARC. It would be intersting to see what Solaris on SPARC could do against linux, assuming they both had similar hardward cost constraints.
Also, I thought an IDE hard drive an odd choice for a high performance server test - wouldn't SCSI or RAID be more appropriate?
Re:Most americans can't take animation seriously
on
End Of Fox Animation
·
· Score: 2
guyver series, nge, macross II, ninja scroll and akira are SERIOUS?
Come on, they are shoot-em-up action flicks. They are good, shoot-em-up action flicks, but they are NOT SERIOUS film
If you want serious film look at films by noted directors, look at films which don't center around a giant special effects spectacular, ninjas, or giant robots.
Imagine a future in which certain Web sites, backed by the power of the law, can control which services can search them.
It exists today - its called hosts.deny.
NO ONE has a right to search your website. If you make it accessible then they can, but every server admin has the right to arbitrarily deny connections. If I want I can stopper everything from *.yahoo.com or *.googlebot.com or even *.uk. We don't need a law to enforce this, it exists at present.
If you were a parent and you know that a 2 ton SUV performs better in a crash test than a half ton econobox, and you rather like the idea of your children surviving a crash, a big SUV suddenly seems like a nice idea and saving a few dollars on gas and worldwide reducing airborne emmissions by 0.0000000000001% seems like a small thing indeed.
Not that I am saying SUVs are the best choice for everyone, its just that a lot of people do have good reasons for getting them.
I believe (like a good libertarian should) that there should be less government. That the government should keep its hands out of just about everything. But on the otherhand, without the government, toysmart can violate my rights.
what sort of libertarian are you?
Libertarianism isn't "no government" or "weak government" it is "controled government"
Libertarians believe that government does have its place and should certainly do its job - its job is to protect our rights and uphold contracts. This toysmart thing perfectly falls underthat - a contract has been violated.
After reading the article I think that the authors are a bit too anxious to discard computers.
True, if used incorrectly computers can be less than worthless in a classroom. But, just because they CAN be misused does not mean we should ban them. Better to research better ways to use them.
Sitting a kid in front of a web browser and shouting 'LEARN!' isn't going to work, most of the edutainment programs are jokes, but at the rate civilization is going basic computer use is going to be as necessary as literacy to survive.
We do need better programs to teach, we can't rely on computers as the sole teacher, but they can, be a useful suppliment.
If you store your belongings in a U-Store it (or any self-service storage place) that doesn't mean that the company can wander in and use your stuff. Just like the stuff you put in a safety deposit box in a bank does not entitle the bank to wander in and take that stuff out.
I am sure that ASPs would have terms of service contracts which will detail these concerns. Otherwise, no sensible company would put their data with a group that provides NO guarnetee of privacy or retrieval.
Perhaps sailors on submarines at war would be a closer example - strict radio silence, high danger.
Or perhaps the sailors of old (such as Magellan's crew) who would go 2 years or so away from port.
Plus, though they may not have voice communication due to have conversations with friends back home, they probably will have enough bandwidth for some good email and slower communications, not to mention if they have a good enough recieved they could still pick up radio and television signals from earth.
This question recently came up on the LAM mailing list (LAM is a free implementation of MPI). and Yes, beowulf is used in a variety of non-academic settings. Many government research labs use it (I am presently working at Lawrence Berkeley on cluster-realted work) and, from the download logs for LAM, I can say that a _lot_ of.com addresses use the sofware. Also, looking over the mailing lists, many people from various large companies use the software.
MPI is an industry standard which has for years been used on Crays, IBMs, HPs and other mainframe and supercomputers. MPI is a highly portable API, so porting to PC clusters can often be just a recompile (realisticly it sometimes takes a bit more tuning, but no where near as bad as say porting from X to Win32 graphics API or something like that)
From the companies' web page it looks like they are powered with MPI, not PVM, not a thread model.
I'd be very intersted in how much of MPI and MPI-2 they support...
What would be intersting is having a number of these cards connected togeteher in the same machine, using MPI for on-card communication, and then some sort of IMPI or some other protcol for communcation to other cards (or other machines)
Why would a company with the smartest people in the world...
Oh here WE go again... do you really think that MS has the smartest people in the world? Or even a large number of them?
Certainly MS has some brilliant people, but there are a lot of very smart people who work elsewhere - sometimes not even in computers (*gasp*!)
I doubt that most microsoft employees, or even a signifigant chunck of them are as smart as say, nobel prize winning physicists, or neruosurgeons, or biologists unlocking the secrets of life, or even most CS professors at major univiersities.
Computers are a big part of today's society, but lets not forget that to to vast vast bulk of the world they are just a tool used to solve problems.
ask yourself, if you were a brilliant engineer or scientist, which would you rather work on - discoverign the nature of the universe, unlocking the cure for deadly diseases, or advancing a word processing program from version 8.0.1 to 8.1?
These are just books and papers I have sitting on my desk Right Now. All of them are advancing fields of research which are being advanced signifigantly in government and university research.
I think his pessimism is overdone.
Most 'new' OSes are just UNIX...on the surface. UNIX is just an interface really. Below the surface there is a lot of innovation there. Look at microkernels, OSes like Plan9 or Inferno...
True, academic research may not produce stunning end-user apps or visual interfaces, but that's because those are by definition marketing-oriented products which take a lot of surveying and preferance finding.
Not to mention that Apple wanted Steve Jobs. Say what you will about him, he inspired the company when it was feeling pretty down and has made it an innovative force again. (again, criticize the iMac all you want, but people like 'em, and after apple starts shipping OS X client, it will be the largest single UNIX distributor on the planet and the only credible non-windows desktop platform out there)
It also would not have been a very polictic way to run a business for Apple to try and support Be. Apple Released specs, 2 open source OSes, and uses a lot of standard components (SCSI, Firewire, IDE, PCI) , but remember that until a few months ago Apple was in serious trouble. Their stock was falling, there were rumors of a buyout, and their OS roadmap was sketchy at best. At this same time Be got a huge chunk of money from Intel. Apple was is no position to send a pack of developers over to help a competitor, Be was angry at Apple for choosing NeXT over BeOS for OS X and had just gotten a huge chunck of change from Intel. Both companies made a reasonable choice - Apple conentrated on making Apples and MacOS, Be switched to the larger Intel Market. They probably would have even if apple did open even more specs just because its a larger market and for a small company like Be they can't afford to be spread thin.
For a while now people have been saying that if we only ran our programs through profilers and then recompiled with the hints they gave we could see a signifigant benifit (in things like branch prediction and such, and other things which cannot be seen at compile-time). I wonder how much more effective having a whole run-time enviorment like this is compared with if the same program used profiling feed back. The run-time enviornment has more up-to-the-second info, but there is the overhead of the enviornment itself, where as profiling would not have that.
what if you want to use emacs, mom wants to watch the DVD, little johnny wants to play games on the play station, and someone else wants to listen to the Internet Radio, all while grandma wants to watch a good old-fashioned VCR tape without having to login and navigate through a dozen menues or wants to sit on the couch and not have to squint at a 17 monitor, but prefers a big fat TV screen?
get each one a computer and that's quite expensive...
But modular set-top boxes for each can be cheaper.
obviously if you have the money, deck your house in a dozen PCs, but if you don't happen to have $10,000 then getting specialized hardware may be the better choice. Also, things like a dreamcast or DVD player may be several hundred now, but they historically come down in price.
the reversal was granted on the basis that Sony should not have been granted the preliminary injunction, not that Connectix's use of the PlayStation BIOS isn't an infringement of Sony's intellectual property.
This stops the preliminary injunction, but not the case, as I understand it.
Re:MacOS Rumors: Not Very Credible
on
Darwin on Crusoe?
·
· Score: 1
Good God! you mean these Mac OS Rumors are _Rumors_ in the sense of Rumors, not Rumors in the sense of verified fact?
I don't think that this would be a 'fork' of OpenGL. OpenGL is an API, not a program, this is just another implementation of it. Mesa is slow and not fully compliant to the official API, SGI's new offering offers the possibility of greater speed and is more compliant. There is already talk of merging it with Mesa to leverage the benifits of both.
Slashdot Mirror
I am not a biochemist either, but I believe that certain types of molecular dynamics simulations and protein folding simulations can be parallized fairly efficently and do not need huge data sets.
Its not a factor of how many operations that need to be performed, its a matter of how many can be performed INDEPENDANTLY of the other processing nodes.
Look at IBM's Blue Gene supercomputer which is using relativley small capacity PIM (processor-in-memory) chips to do large protien folding calculations.
BTW, IANAB (I am not a biologist), so I can't say exactly what sort of simulations they are doing...
Who cares how many servers are deployed? It is irrelevant to EVERYONE. Your decision to use an OS should be completely independent of what other people are doing, right?
Idealisticly, yes. Realisticly, NO.
If I use an OS (or any product) which is commonly used then I will have a better chance of knowing that it will be supported by more vendors, more developers will make programs for it, and there will be a support community for it in the future.
Knowing the usage statistics for a platoform are important in that they allow you to know "if I go with product X, will I be stranded and forced to migrate in 3 years? or will there be a growing community developing and advancing the product?"
The quote that "the idea is to get rid of the keyboard" did _not_ come from Apple. It came from a source "familar with Apple's plans" which could mean just about nothing.
I doubt that apple would scrap keyboards on all of their powerbooks, at most they may offer one or two models without and add handwriting recognition to the rest.
It looks like they used Solaris x86, which is not as finely tuned as Solaris/SPARC. It would be intersting to see what Solaris on SPARC could do against linux, assuming they both had similar hardward cost constraints.
Also, I thought an IDE hard drive an odd choice for a high performance server test - wouldn't SCSI or RAID be more appropriate?
guyver series, nge, macross II, ninja scroll and akira are SERIOUS?
Come on, they are shoot-em-up action flicks. They are good, shoot-em-up action flicks, but they are NOT SERIOUS film
If you want serious film look at films by noted directors, look at films which don't center around a giant special effects spectacular, ninjas, or giant robots.
I love anime, but I wouldn't call it "serious".
Wait, isn't Tau Neutrino Amiga's new OS????
It exists today - its called hosts.deny.
NO ONE has a right to search your website. If you make it accessible then they can, but every server admin has the right to arbitrarily deny connections. If I want I can stopper everything from *.yahoo.com or *.googlebot.com or even *.uk. We don't need a law to enforce this, it exists at present.
If you were a parent and you know that a 2 ton SUV performs better in a crash test than a half ton econobox, and you rather like the idea of your children surviving a crash, a big SUV suddenly seems like a nice idea and saving a few dollars on gas and worldwide reducing airborne emmissions by 0.0000000000001% seems like a small thing indeed.
Not that I am saying SUVs are the best choice for everyone, its just that a lot of people do have good reasons for getting them.
I believe (like a good libertarian should) that there should be less government. That the government should keep its hands out of just about everything. But on the otherhand, without the government, toysmart can violate my rights.
what sort of libertarian are you?
Libertarianism isn't "no government" or "weak government" it is "controled government"
Libertarians believe that government does have its place and should certainly do its job - its job is to protect our rights and uphold contracts. This toysmart thing perfectly falls underthat - a contract has been violated.
Fortunately, this could never happen here
I think he was being sarcastic.
After reading the article I think that the authors are a bit too anxious to discard computers.
True, if used incorrectly computers can be less than worthless in a classroom. But, just because they CAN be misused does not mean we should ban them. Better to research better ways to use them.
Sitting a kid in front of a web browser and shouting 'LEARN!' isn't going to work, most of the edutainment programs are jokes, but at the rate civilization is going basic computer use is going to be as necessary as literacy to survive.
We do need better programs to teach, we can't rely on computers as the sole teacher, but they can, be a useful suppliment.
In all honesty, I'm more interested in scans of the bootie chicks, propaganda I get enough of as is.
If you store your belongings in a U-Store it (or any self-service storage place) that doesn't mean that the company can wander in and use your stuff. Just like the stuff you put in a safety deposit box in a bank does not entitle the bank to wander in and take that stuff out.
I am sure that ASPs would have terms of service contracts which will detail these concerns. Otherwise, no sensible company would put their data with a group that provides NO guarnetee of privacy or retrieval.
Perhaps sailors on submarines at war would be a closer example - strict radio silence, high danger.
Or perhaps the sailors of old (such as Magellan's crew) who would go 2 years or so away from port.
Plus, though they may not have voice communication due to have conversations with friends back home, they probably will have enough bandwidth for some good email and slower communications, not to mention if they have a good enough recieved they could still pick up radio and television signals from earth.
I think they'll make it.
This question recently came up on the LAM mailing list (LAM is a free implementation of MPI). and Yes, beowulf is used in a variety of non-academic settings. Many government research labs use it (I am presently working at Lawrence Berkeley on cluster-realted work) .com addresses use the sofware. Also, looking over the mailing lists, many people from various large companies use the software.
and, from the download logs for
LAM, I can say that a _lot_ of
MPI is an industry standard which has for years been used on Crays, IBMs, HPs and other mainframe and supercomputers. MPI is a highly portable API, so porting to PC clusters can often be just a recompile (realisticly it sometimes takes a bit more tuning, but no where near as bad as say porting from X to Win32 graphics API or something like that)
From the companies' web page it looks like they are powered with MPI, not PVM, not a thread model.
I'd be very intersted in how much of MPI and MPI-2 they support...
What would be intersting is having a number of these cards connected togeteher in the same machine, using MPI for on-card communication, and then some sort of IMPI or some other protcol for communcation to other cards (or other machines)
Why would a company with the smartest people in the world...
Oh here WE go again... do you really think that MS has the smartest people in the world? Or even a large number of them?
Certainly MS has some brilliant people, but there are a lot of very smart people who work elsewhere - sometimes not even in computers (*gasp*!)
I doubt that most microsoft employees, or even a signifigant chunck of them are as smart as say, nobel prize winning physicists, or neruosurgeons, or biologists unlocking the secrets of life, or even most CS professors at major univiersities.
Computers are a big part of today's society, but lets not forget that to to vast vast bulk of the world they are just a tool used to solve problems.
ask yourself, if you were a brilliant engineer or scientist, which would you rather work on - discoverign the nature of the universe, unlocking the cure for deadly diseases, or advancing a word processing program from version 8.0.1 to 8.1?
Generic Programming (STL & others), pervasive threading (HEP, Monsoon), Standardized C++, Processor-in-memory architechture, Quantum Computing (QCA, true quantum computing), Clustering (Beowulf), Message Passing (MPI), OpenMP, Virtual Interface Archtechture networking...
These are just books and papers I have sitting on my desk Right Now. All of them are advancing fields of research which are being advanced signifigantly in government and university research.
I think his pessimism is overdone.
Most 'new' OSes are just UNIX...on the surface. UNIX is just an interface really. Below the surface there is a lot of innovation there. Look at microkernels, OSes like Plan9 or Inferno...
True, academic research may not produce stunning end-user apps or visual interfaces, but that's because those are by definition marketing-oriented products which take a lot of surveying and preferance finding.
Not to mention that Apple wanted Steve Jobs. Say what you will about him, he inspired the company when it was feeling pretty down and has made it an innovative force again. (again, criticize the iMac all you want, but people like 'em, and after apple starts shipping OS X client, it will be the largest single UNIX distributor on the planet and the only credible non-windows desktop platform out there)
It also would not have been a very polictic way to run a business for Apple to try and support Be. Apple Released specs, 2 open source OSes, and uses a lot of standard components (SCSI, Firewire, IDE, PCI) , but remember that until a few months ago Apple was in serious trouble. Their stock was falling, there were rumors of a buyout, and their OS roadmap was sketchy at best. At this same time Be got a huge chunk of money from Intel. Apple was is no position to send a pack of developers over to help a competitor, Be was angry at Apple for choosing NeXT over BeOS for OS X and had just gotten a huge chunck of change from Intel. Both companies made a reasonable choice - Apple conentrated on making Apples and MacOS, Be switched to the larger Intel Market. They probably would have even if apple did open even more specs just because its a larger market and for a small company like Be they can't afford to be spread thin.
For a while now people have been saying that if we only ran our programs through profilers and then recompiled with the hints they gave we could see a signifigant benifit (in things like branch prediction and such, and other things which cannot be seen at compile-time). I wonder how much more effective having a whole run-time enviorment like this is compared with if the same program used profiling feed back. The run-time enviornment has more up-to-the-second info, but there is the overhead of the enviornment itself, where as profiling would not have that.
what if you want to use emacs, mom wants to watch the DVD, little johnny wants to play games on the play station, and someone else wants to listen to the Internet Radio, all while grandma wants to watch a good old-fashioned VCR tape without having to login and navigate through a dozen menues or wants to sit on the couch and not have to squint at a 17 monitor, but prefers a big fat TV screen?
get each one a computer and that's quite expensive...
But modular set-top boxes for each can be cheaper.
obviously if you have the money, deck your house in a dozen PCs, but if you don't happen to have $10,000 then getting specialized hardware may be the better choice. Also, things like a dreamcast or DVD player may be several hundred now, but they historically come down in price.
check out this article on the register.
the reversal was granted on the basis that Sony should not have
been granted the preliminary injunction, not that Connectix's use of the
PlayStation BIOS isn't an infringement of Sony's intellectual property.
This stops the preliminary injunction, but not the case, as I understand it.
Good God!
you mean these Mac OS Rumors are _Rumors_ in the sense of Rumors, not Rumors in the sense of verified fact?
Ohh... I hate that.
I don't think that this would be a 'fork' of OpenGL. OpenGL is an API, not a program, this is just another implementation of it. Mesa is slow and not fully compliant to the official API, SGI's new offering offers the possibility of greater speed and is more compliant. There is already talk of merging it with Mesa to leverage the benifits of both.