Drilling through the article with my utterly minimal norwegian (Prarie Home Companion + German + exposure to Danish coworkers), I think I've distilled the following:
Cache on the LSI RAID controller is 1/2 the adaptec. Performance is comparable, though not equivalent.
All of the controllers are 64-bit.
Adaptec SCSI is good for both hardware RAID and software RAID.
LSI has good hardware SCSI RAID only.
Don't use current SATA controllers (RAID or Otherwise) for best performance.
Does anybody with access to a good collection of modern hardware care to re-run this test in a language that Babelfish understands?
I built my penultimate cluster from IBM 1U opteron systems, using Rocks (http://www.rocksclusters.org/ as the system software. That was a decent setup from a price-performance standpoint, ran well, and supported our legacy 32-bit linux apps rather well. Currently I'm using Apple XServe G5s, mostly because of the availability of XLF and the low power-draw. I also don't have the legacy binary issues in the new lab that I had when supporting a heterogeneous group.
Having started with white-boxes on bread-racks, passed through IA-64 and SP2, and ended up with racked 1U, I would say that what I care about is reliability and support first, speed second, and convenient software environment third. Google can afford to have a system fail, due to their distributed structure and design for failover, but tightly-coupled scientific codes tend not to tolerate this as well.
Therefore, my recommendation would be for scientific clustering, (2-way proc, 1U/2U rack enclosure from IBM, HP, Apple), G5/Opteron > Xeon. If I had the money, then swap G5/Opteron for Power4/IA-64(HP). Rack mounting saves more than just space; it's also easier to perform maintenance on than the bread-racked desktops had been.
Note, my opinion only applies to my environment; high-performance/high-throughput computing. High availability clusters are another beast entirely.
Which is purchased depends on who is in charge; the geeks or the jocks. Remember High School? Same story: UCAVs are cheap, stealthy, effective, and flown by people with good hand-eye coordination from armchairs somewhere, while fighter planes are expensive, flamboyant, effective, and require macho fighter jocks to pilot them.
Militarily, UCAVs seem the way to go, but in practice, there's no glory, promotion, or machismo in using them. You'll have to dismantle the entire approximately 100 year old culture of the Flying Ace to get them through the Air Force in large numbers. Your best hope is turning them over to the Army and Marines, who have a ground-pounder culture, and will fight for funding for UCAVs to support their troops while choking off the funding for junior birdmen over in the air force.
It's nice to see Pascal-family languages attempt a comeback against the Line-Noise-Family (you know who you are).
I started with UCSD P-System (yuck), moved to TurboPascal 2.0 (nice), and ended up with VAX Pascal (marvelous). The only issue was that with 6 - 8 comp. sci. projects being compiled on an 11/750, it got terribly slow, but the strings weren't compatible between Turbo-P and VAX-P. I seem to remember writing an alternate STRING datatype laid out like the VAX expected it, so that I could prototype on my QX-16, then upload to the VAX for final compilation. Turbo-Pascal using the Wordstar key sequences helped immensely as well.
Everything old is new again; FreePascal hits 2.0, GFortran includes F95 (which is enough like non-OO Pascal/Ada for the desperate), and Smalltalk (Squeak) is being used for large software. Now all we need is GNU-Algol.
In some cases, yes, but they're all much more expensive than the disposable alternatives.
Look at the most obvious example of why it's not necessarily a conspiracy keeping these fine products away from you; your computer. IBM, HP, and DEC made some high-end, virtually unkillable, PCs (and two of them still do), but they cost real money. People said that was too much money for a PC, so they bought Dell or Bob's WhiteBoxen instead, and complained when they died early.
It's not that the company loses future sales by not selling you an undullable razor for $50 now, it's that they lose present sales because Bob's FaceScrapers(tm) are only $2.50 at the check-out counter, and that's what the average short-term minded consumer will buy instead. You could use compact fluorescent bulbs, or LED flashlights as examples as well; some people will buy them, but most will say it's too expensive, and instead pony up for another tungsten bulb seven times as often, because it costs less at that moment.
Trite though the statement is, the business world really does run on the maxim, "never attribute to malice what can be adequately explained by stupidity".
Yes, it will be sad if Europe reverts to the temperature range of Canada or Russia. (French Ice Wine, anyone?) On the other hand, NJ and much of the east coast is also warmer than it should be due to the Gulf Stream.
No more people moving from NY/VT/NH to Florida, etc., for the climate and ruining our tax base!
Well, is anyone sure that there's a problem here: either the paper was valid, and rejected by an editor for conflict of interest, in which case there is a problem, but the paper should have gone to another journal, or it's not of interest, and was correctly bounced. It seems a valid test would be to submit it to another journal, and see if they bite.
Going the advertisement/straight to the media route is generally bad, as that is what bit Pons and Fleischman.
A more interesting question would be why not, in the age of web-based publishing, don't scientific journals adopt some variant of/.'s rating system? ACS pre-publishes articles on the web (i.e. available before the journal issue has the pages ready), so in theory you could allow authorized users to rate them. I'd love to be able to browse J. Phys. Chem. A. (for instance) at +2, so as to quickly find which articles my colleagues find interesting, then go back and give the rest of the articles a closer read later.
I always thought Desktop Linux was a communist plot to get you to spend all of your time tweaking your desktop settings (Because You Can!), rather than doing actual work.
Now we just have to find out who's plot making Windows users reboot all the time was.
Good idea. You've upgraded them from a dremel tool to a chainsaw. Now when they get "owned" (I refuse to keep up on the weekly l337 spelling changes), they can start to do real damage, rather than just sending more spam.
The only way this is a safer solution is that because they'll be continually recompiling their entire OS (granting agencies just love the excuse, "I missed the deadline because I'm recompiling my word processor"), they'll never have any time to actually do something destructive with their computer.
I was a kid of about 7 or 8 when TOS started re-running in our area, and already interested in science. I seem to remember being thrilled about Scotty and Spock; i.e. the people interested in science and technology were visible, important, and accepted.
I remember part of the charm when I was a bit older was that it posited a future where with some combination of technology, science, democracy, human rights, and the guts to stand up for them, a better world was possible.
Yes, it's bad acting, half-baked plots, and a cult that I don't pretend to understand, but for 8 year old nerds everywhere, it did offer a positive vision.
Having had my quotes "ahem" edited by a Large Blue Corporation's advertising team, it's easy to see how some of the switcher ads sounded like they were purely made up by marketing. In reality, someone took out the "ums" and "errs" (except in Ellen Feiss's case), and tightened the grammar in the process.
The Mac switchers were probably just incoheret from relief.
Does C++ now? I remember years of programming examples on how to implement a complex type into C++, which then ignored the problem that Fortran also had exponentiation, multiple-precisions, etc, already defined.
I'm sorry (truly) that Ada didn't catch on like it should have, as I was trained in VAX Pascal, and am still more comfortable with Pascal-family languages. It's especially strange when you consider that Ada's syntax and F90's are remarkably similar.
Maybe the problem is that the language has to be named Fortran before scientists will use it?
Specifically, it's F77 -> F90 -> F95 -> F2K. There have been enough attempts to replace Fortran, and the only result so far is that it's kept computer scientists entertained. All of these ideas are driven by one common thread; formally trained computer scientists can't stand Fortran 77's control structures, non-dynamic memory, etc, and demand that it must be replaced for religious reasons. F90/F95 have already fixed those problems, but it's still called Fortran, and so it simply *MUST* be replaced.
Let's see, we had PL/I (a merger of Fortran, COBOL, and Algol), RATFOR, Ada, Matlab, C++, and the late, and rather lamented, Sather. None of them has the performance of Fortran, the ease of programming, the extensive and validated libraries, complex numbers as a fundamental data type, or the solidity of compilers.
It's the cockroach of computer languages; you can keep spraying, and it will keep sneakout out at night.
As much as I favor family unity, supporting one' other half, etc., I would have to say where childbirth comes in, I'd suggest the movie. I'd further suggest it in another city. Not all of the old ways, (father pacing outside, mother and doctors inside somewhere), were all that bad.
Plus, as Rita Rudner once put it, "If my husband think's he's going to 'share the experience', I expect him to be tied to the table next to me, getting his legs waxed."
The system of putting everything in one folder is called a "minidisk", and was still common on IBM VM/XA and VM/CMS systems a few years ago. Our IBM rep used to try to convince us that virtual piles of unnested minidisks were an ideal way to manage our files. After all, it worked so well on our physical desktops.
Weight may not be a factor, but flexibility is. Traditional superconductors are ceramics, where breaks between domains ruins the transmission. Carbon nanotubes, OTOH, would be flexible, and could be routed in manners than relatively rigids ceramics couldn't. The would also be more resistant to failures due to flexing.
It would be interesting to know the weight of the wire in current launch vehicles, as every kilo less of copper wire is a kilo more of payload you can lob into orbit.
The reason these devices are interesting is the flow of Neutrons.
There are several applications in materials science where you want neutrons, but you don't want to send your sample off to Oak Ridge, and wait, or go through the paperwork to try to build a research reactor. This device would allow, for instance, in-house Neutron Diffraction experiments, which is similar to X-ray diffraction except that Hydrogens show up. You can see hydrogen loading in containment materials, migration in batteries, and other minor structural changes which are invisible to other analytic techniques.
The fact that they use fusion is nifty, but it's the neutron flux in a convenient package that makes this a way cool experiment.
It would be nice to have this followed up. An interesting friday science article in the Wall Street Journal commented on how suspects could be browbeaten into confessing (and believing in) a crime they didn't commit. It would be interesting to investigate whether you could tell the difference between actually did the crime, and merely has been convinced they did the crime.
A little far fetched, but it's the future you're posting from.
"A gram is better than a damn..." Look outside at your productivity medicated, caffeine-slurping, entertainment anaesthetized, coworkers, and try to tell me that Huxley was wrong.
We'll get the decanting right in a few years, though I doubt that Ford worship is on the horizon. (Toyotas, maybe, but not Fords)
Ah, the joy of handing over as evidence a nice, compact, glittering, crystal while telling your superior that your arch-nemesis was only working for himself. Not quite the same with a USB pen-drive and a couple of floppies.
I was called in to fix coworkers blowing away/dev on a new Sparcstation once. I told them they could restore it from/dev/cdrom,/dev/tape, or type it in manually from/dev/keyboard.
Slashdot Mirror
Drilling through the article with my utterly minimal norwegian (Prarie Home Companion + German + exposure to Danish coworkers), I think I've distilled the following:
Cache on the LSI RAID controller is 1/2 the adaptec. Performance is comparable, though not equivalent.
All of the controllers are 64-bit.
Adaptec SCSI is good for both hardware RAID and software RAID.
LSI has good hardware SCSI RAID only.
Don't use current SATA controllers (RAID or Otherwise) for best performance.
Does anybody with access to a good collection of modern hardware care to re-run this test in a language that Babelfish understands?
I built my penultimate cluster from IBM 1U opteron systems, using Rocks (http://www.rocksclusters.org/ as the system software. That was a decent setup from a price-performance standpoint, ran well, and supported our legacy 32-bit linux apps rather well. Currently I'm using Apple XServe G5s, mostly because of the availability of XLF and the low power-draw. I also don't have the legacy binary issues in the new lab that I had when supporting a heterogeneous group.
Having started with white-boxes on bread-racks, passed through IA-64 and SP2, and ended up with racked 1U, I would say that what I care about is reliability and support first, speed second, and convenient software environment third. Google can afford to have a system fail, due to their distributed structure and design for failover, but tightly-coupled scientific codes tend not to tolerate this as well.
Therefore, my recommendation would be for scientific clustering, (2-way proc, 1U/2U rack enclosure from IBM, HP, Apple), G5/Opteron > Xeon. If I had the money, then swap G5/Opteron for Power4/IA-64(HP). Rack mounting saves more than just space; it's also easier to perform maintenance on than the bread-racked desktops had been.
Note, my opinion only applies to my environment; high-performance/high-throughput computing. High availability clusters are another beast entirely.
Which is purchased depends on who is in charge; the geeks or the jocks. Remember High School? Same story: UCAVs are cheap, stealthy, effective, and flown by people with good hand-eye coordination from armchairs somewhere, while fighter planes are expensive, flamboyant, effective, and require macho fighter jocks to pilot them.
Militarily, UCAVs seem the way to go, but in practice, there's no glory, promotion, or machismo in using them. You'll have to dismantle the entire approximately 100 year old culture of the Flying Ace to get them through the Air Force in large numbers. Your best hope is turning them over to the Army and Marines, who have a ground-pounder culture, and will fight for funding for UCAVs to support their troops while choking off the funding for junior birdmen over in the air force.
It's nice to see Pascal-family languages attempt a comeback against the Line-Noise-Family (you know who you are).
I started with UCSD P-System (yuck), moved to TurboPascal 2.0 (nice), and ended up with VAX Pascal (marvelous). The only issue was that with 6 - 8 comp. sci. projects being compiled on an 11/750, it got terribly slow, but the strings weren't compatible between Turbo-P and VAX-P. I seem to remember writing an alternate STRING datatype laid out like the VAX expected it, so that I could prototype on my QX-16, then upload to the VAX for final compilation. Turbo-Pascal using the Wordstar key sequences helped immensely as well.
Everything old is new again; FreePascal hits 2.0, GFortran includes F95 (which is enough like non-OO Pascal/Ada for the desperate), and Smalltalk (Squeak) is being used for large software. Now all we need is GNU-Algol.
In some cases, yes, but they're all much more expensive than the disposable alternatives.
Look at the most obvious example of why it's not necessarily a conspiracy keeping these fine products away from you; your computer. IBM, HP, and DEC made some high-end, virtually unkillable, PCs (and two of them still do), but they cost real money. People said that was too much money for a PC, so they bought Dell or Bob's WhiteBoxen instead, and complained when they died early.
It's not that the company loses future sales by not selling you an undullable razor for $50 now, it's that they lose present sales because Bob's FaceScrapers(tm) are only $2.50 at the check-out counter, and that's what the average short-term minded consumer will buy instead. You could use compact fluorescent bulbs, or LED flashlights as examples as well; some people will buy them, but most will say it's too expensive, and instead pony up for another tungsten bulb seven times as often, because it costs less at that moment.
Trite though the statement is, the business world really does run on the maxim, "never attribute to malice what can be adequately explained by stupidity".
Those are graduate Students: their life (and future career) *DOES* depend on their running towards the snake.
If it gets away, guess who gets to go back out into the jungle to catch a new one?
To the reader suggesting farming them, that's how the problem got started.
Now the Nutria are such a problem that they once paid Paul Prudhomme to come up with Nutria recipes, to encourage people to go out and eat them.
Based on my last trip to Dixie Kitchen before moving east, I don't think those recipes caught on.
Yes, it will be sad if Europe reverts to the temperature range of Canada or Russia. (French Ice Wine, anyone?) On the other hand, NJ and much of the east coast is also warmer than it should be due to the Gulf Stream.
No more people moving from NY/VT/NH to Florida, etc., for the climate and ruining our tax base!
Well, is anyone sure that there's a problem here: either the paper was valid, and rejected by an editor for conflict of interest, in which case there is a problem, but the paper should have gone to another journal, or it's not of interest, and was correctly bounced. It seems a valid test would be to submit it to another journal, and see if they bite.
/.'s rating system? ACS pre-publishes articles on the web (i.e. available before the journal issue has the pages ready), so in theory you could allow authorized users to rate them. I'd love to be able to browse J. Phys. Chem. A. (for instance) at +2, so as to quickly find which articles my colleagues find interesting, then go back and give the rest of the articles a closer read later.
Going the advertisement/straight to the media route is generally bad, as that is what bit Pons and Fleischman.
A more interesting question would be why not, in the age of web-based publishing, don't scientific journals adopt some variant of
I always thought Desktop Linux was a communist plot to get you to spend all of your time tweaking your desktop settings (Because You Can!), rather than doing actual work.
Now we just have to find out who's plot making Windows users reboot all the time was.
Good idea. You've upgraded them from a dremel tool to a chainsaw. Now when they get "owned" (I refuse to keep up on the weekly l337 spelling changes), they can start to do real damage, rather than just sending more spam.
The only way this is a safer solution is that because they'll be continually recompiling their entire OS (granting agencies just love the excuse, "I missed the deadline because I'm recompiling my word processor"), they'll never have any time to actually do something destructive with their computer.
I was a kid of about 7 or 8 when TOS started re-running in our area, and already interested in science. I seem to remember being thrilled about Scotty and Spock; i.e. the people interested in science and technology were visible, important, and accepted.
I remember part of the charm when I was a bit older was that it posited a future where with some combination of technology, science, democracy, human rights, and the guts to stand up for them, a better world was possible.
Yes, it's bad acting, half-baked plots, and a cult that I don't pretend to understand, but for 8 year old nerds everywhere, it did offer a positive vision.
Having had my quotes "ahem" edited by a Large Blue Corporation's advertising team, it's easy to see how some of the switcher ads sounded like they were purely made up by marketing. In reality, someone took out the "ums" and "errs" (except in Ellen Feiss's case), and tightened the grammar in the process.
The Mac switchers were probably just incoheret from relief.
Does C++ now? I remember years of programming examples on how to implement a complex type into C++, which then ignored the problem that Fortran also had exponentiation, multiple-precisions, etc, already defined.
I'm sorry (truly) that Ada didn't catch on like it should have, as I was trained in VAX Pascal, and am still more comfortable with Pascal-family languages. It's especially strange when you consider that Ada's syntax and F90's are remarkably similar.
Maybe the problem is that the language has to be named Fortran before scientists will use it?
It's probably under the BSD license. Now you have to have four paragraphs of advertising tatooed onto your chest.
The successor to Fortran, is Fortran.
Specifically, it's F77 -> F90 -> F95 -> F2K. There have been enough attempts to replace Fortran, and the only result so far is that it's kept computer scientists entertained. All of these ideas are driven by one common thread; formally trained computer scientists can't stand Fortran 77's control structures, non-dynamic memory, etc, and demand that it must be replaced for religious reasons. F90/F95 have already fixed those problems, but it's still called Fortran, and so it simply *MUST* be replaced.
Let's see, we had PL/I (a merger of Fortran, COBOL, and Algol), RATFOR, Ada, Matlab, C++, and the late, and rather lamented, Sather. None of them has the performance of Fortran, the ease of programming, the extensive and validated libraries, complex numbers as a fundamental data type, or the solidity of compilers.
It's the cockroach of computer languages; you can keep spraying, and it will keep sneakout out at night.
As much as I favor family unity, supporting one' other half, etc., I would have to say where childbirth comes in, I'd suggest the movie. I'd further suggest it in another city. Not all of the old ways, (father pacing outside, mother and doctors inside somewhere), were all that bad.
Plus, as Rita Rudner once put it, "If my husband think's he's going to 'share the experience', I expect him to be tied to the table next to me, getting his legs waxed."
The system of putting everything in one folder is called a "minidisk", and was still common on IBM VM/XA and VM/CMS systems a few years ago. Our IBM rep used to try to convince us that virtual piles of unnested minidisks were an ideal way to manage our files. After all, it worked so well on our physical desktops.
Weight may not be a factor, but flexibility is. Traditional superconductors are ceramics, where breaks between domains ruins the transmission. Carbon nanotubes, OTOH, would be flexible, and could be routed in manners than relatively rigids ceramics couldn't. The would also be more resistant to failures due to flexing.
It would be interesting to know the weight of the wire in current launch vehicles, as every kilo less of copper wire is a kilo more of payload you can lob into orbit.
The reason these devices are interesting is the flow of Neutrons.
There are several applications in materials science where you want neutrons, but you don't want to send your sample off to Oak Ridge, and wait, or go through the paperwork to try to build a research reactor. This device would allow, for instance, in-house Neutron Diffraction experiments, which is similar to X-ray diffraction except that Hydrogens show up. You can see hydrogen loading in containment materials, migration in batteries, and other minor structural changes which are invisible to other analytic techniques.
The fact that they use fusion is nifty, but it's the neutron flux in a convenient package that makes this a way cool experiment.
It would be nice to have this followed up. An interesting friday science article in the Wall Street Journal commented on how suspects could be browbeaten into confessing (and believing in) a crime they didn't commit. It would be interesting to investigate whether you could tell the difference between actually did the crime, and merely has been convinced they did the crime.
A little far fetched, but it's the future you're posting from.
"A gram is better than a damn..." Look outside at your productivity medicated, caffeine-slurping, entertainment anaesthetized, coworkers, and try to tell me that Huxley was wrong.
We'll get the decanting right in a few years, though I doubt that Ford worship is on the horizon. (Toyotas, maybe, but not Fords)
Ah, the joy of handing over as evidence a nice, compact, glittering, crystal while telling your superior that your arch-nemesis was only working for himself. Not quite the same with a USB pen-drive and a couple of floppies.
I was called in to fix coworkers blowing away /dev on a new Sparcstation once. I told them they could restore it from /dev/cdrom, /dev/tape, or type it in manually from /dev/keyboard.
An SP2 should be blazingly fast. Just get the HIPPI interconnect, upgrade to Nighthawk nodes, and install AIX 5.
Oh wait, Windows SP2. Nothing to be done for that. Sorry.