I would be interested to find
out what portion of the cost is attributed to communication, but I wouldn't be surprized if it were most of the
money.
Only a minority of the cost is Myrinet. People tend to think it's far more expensive than it actually is. If we had used Intel or AMD processors, and still had 1 gigabit of bandwidth per processor, then Myrinet would be more but still not a majority.
Lots of government labs are running Beowulfs. But the press release says that this is the first "turn key" Linux-based supercomputer. That's quite a difference.
This contract includes 2 substantial upgrades; this is just the initial installation. The AlphaLinux cluster (yes, connected with Myrinet) is most of the initial equipment. There's also a tape robot from ADIC with 70 terabytes of tape (1400 tapes) and 20 tape drives, and a storage area network (SAN) using CVFS, a SAN filesystem being ported to Linux because of this contract.
The main software used on the system is actually all free: Linux, the PBS batch queue system, mpich as modified by Myricom for MPI, and the SMS scalable modeling system, developed at FSL. FSL has demonstrated some of their software scaling efficiently up to around 100 nodes. Limits in scalablility, the Alpha's superior floating point performance, and Compaq's great AlphaLinux compilers are the reason we used Alphas.
The only thing that really competes with cray on the highend is the SGIO2k...hmmm, wonder where they got that technology?
Uh, SGI developed the O2k before they bought Cray, and there is lots of competition on the high end. For example, for the FSL weather forecasting bid, SGI bid an O2k, but the winner was a cluster running AlphaLinux.
For a wind tunnel simulation you can assign different sections of the tunnel to different processors and each one communicates only with its neighbors with nice, predictable access patters. For something like a simulation of a car crash this will not work too well, though.
No, car crashes are an example where clusters do just fine. While the Tera does do a bit better with load balancing on codes like that, cluster hardware is so cheap that you can just buy lots more cluster, and not worry much about load balance. That's the problem that has already made Tera's MTA not succeed in the market; it's not cost effective for any problem.
Intel dropped ot of the supercomputer market a whlie ago, right after they produced ASCI Red. Sure, their processors are used by other companies to build supercomputers, but the same is true for the Compaq Alpha, and soon for AMD's Athlon.
There are some new kids on the supercomputer block, Linux-using kids, even. Too bad Slashdot gives the old, stodgy, non-open-source supercomputing companies lots more press;-)
A 1 GHz Pentium is not a supercomputer -- a supercomputer is a moving target. Today's supercomputers get a few hundred gigaflops peak performance, not less tha one.
First off, ESR didn't invent the difference between "cracker" and "hacker" -- a lot of the community also divides it that way, far more than just people who actually care what ESR says.
Second, if security people are actually complaining to journalists about use of the word "crack", they ought to get a clue. "I cracked a password", "I cracked the copy protection on that software", and "I cracked that site" all have obvious meanings. Hint: the 3rd doesn't mean that I broke some passwords that I could use to break into the site, but didn't actually break in.
We shouldn't spam anyone. On the other hand, Clemson is a public university, so a lot more people than just the students have a stake in what Clemson does.
The E10k is what I beat. My customer isn't thankful that the application runs on a cluster; it was designed to run on a cluster. And operating system has a LOT to do with it; we have customized several things in the kernel, which we couldn't do in Solaris.
It's funny that Sun says that Linux is "low end", while my company beat Sun (and IBM and HP and SGI) on a $15 million dollar supercomputer bid with a cluster running... Linux.
I guess that's just too little money for Sun to be interested.
0.17 micro is small? RAM chips are usually at least 1 process generation ahead of CPUs. You should be asking why they aren't using 0.13 micron for this.
Samsung is currently sampling Alphas in 0.18 micron.
Linux does have to compete with other Unixes, but people often decide in Linux's favor. For example, this cluster is 277 nodes with better networking, and we chose Linux over Tru64, due to Linux's super system administration capabilities.
BTW, you can get Compaq's great Alpha compilers for Linux.
I find it funny that no one challenged Tom's claim that Unisys invested resources to invent LZW.
LZW was invented by two guys in Israel, Lempel and Ziv. Unisys (and Welch) patented their work, and doesn't give them a dime. It's completely legal, yet stupid. This case has always been the poster child of the "Patents are Lame" movement.
Whenever you see Unisys whining about all the money they spent to develop LZW, remember the facts.
Slashdot Mirror
Only a minority of the cost is Myrinet. People tend to think it's far more expensive than it actually is. If we had used Intel or AMD processors, and still had 1 gigabit of bandwidth per processor, then Myrinet would be more but still not a majority.
Lots of government labs are running Beowulfs. But the press release says that this is the first "turn key" Linux-based supercomputer. That's quite a difference.
This contract includes 2 substantial upgrades; this is just the initial installation. The AlphaLinux cluster (yes, connected with Myrinet) is most of the initial equipment. There's also a tape robot from ADIC with 70 terabytes of tape (1400 tapes) and 20 tape drives, and a storage area network (SAN) using CVFS, a SAN filesystem being ported to Linux because of this contract.
The main software used on the system is actually all free: Linux, the PBS batch queue system, mpich as modified by Myricom for MPI, and the SMS scalable modeling system, developed at FSL. FSL has demonstrated some of their software scaling efficiently up to around 100 nodes. Limits in scalablility, the Alpha's superior floating point performance, and Compaq's great AlphaLinux compilers are the reason we used Alphas.
Uh, SGI developed the O2k before they bought Cray, and there is lots of competition on the high end. For example, for the FSL weather forecasting bid, SGI bid an O2k, but the winner was a cluster running AlphaLinux.
No, car crashes are an example where clusters do just fine. While the Tera does do a bit better with load balancing on codes like that, cluster hardware is so cheap that you can just buy lots more cluster, and not worry much about load balance. That's the problem that has already made Tera's MTA not succeed in the market; it's not cost effective for any problem.
Intel dropped ot of the supercomputer market a whlie ago, right after they produced ASCI Red. Sure, their processors are used by other companies to build supercomputers, but the same is true for the Compaq Alpha, and soon for AMD's Athlon.
There are some new kids on the supercomputer block, Linux-using kids, even. Too bad Slashdot gives the old, stodgy, non-open-source supercomputing companies lots more press
A 1 GHz Pentium is not a supercomputer -- a supercomputer is a moving target. Today's supercomputers get a few hundred gigaflops peak performance, not less tha one.
First off, ESR didn't invent the difference between "cracker" and "hacker" -- a lot of the community also divides it that way, far more than just people who actually care what ESR says.
Second, if security people are actually complaining to journalists about use of the word "crack", they ought to get a clue. "I cracked a password", "I cracked the copy protection on that software", and "I cracked that site" all have obvious meanings. Hint: the 3rd doesn't mean that I broke some passwords that I could use to break into the site, but didn't actually break in.
Geez.
We shouldn't spam anyone. On the other hand, Clemson is a public university, so a lot more people than just the students have a stake in what Clemson does.
The E10k is what I beat. My customer isn't thankful that the application runs on a cluster; it was designed to run on a cluster. And operating system has a LOT to do with it; we have customized several things in the kernel, which we couldn't do in Solaris.
It's funny that Sun says that Linux is "low end", while my company beat Sun (and IBM and HP and SGI) on a $15 million dollar supercomputer bid with a cluster running... Linux.
I guess that's just too little money for Sun to be interested.
0.17 micro is small? RAM chips are usually at least 1 process generation ahead of CPUs. You should be asking why they aren't using 0.13 micron for this.
Samsung is currently sampling Alphas in 0.18 micron.
Your associate professor is one of many people who don't realize that Fortran has evolved a bit since Fortran 77.
My employer would never have won the $15 million FSL bid
(see this press release) without a good Fortran 95 compiler.
Most people who don't use fortran don't think that anyone uses Fortran. They are mistaken.
Oh great, Slashdot ate my less than sign.
You can get a sufficient quality PC for a scientific cluster for far less than $20k. I'd bet that they don't want 100 raid controllers.
MRI grants do not allow universities to charge overhead, and is 100% hardware money. You also have to get at least 20% matching funds.
In general, equipment over $500 isn't assessed overhead by any university.
MRI grants can't include personnel money or electricity.
... these machines ARE massively parallel supercomputers, if you build them big enough and you use the best commodity networking (like myrinet).
Linux does have to compete with other Unixes, but people often decide in Linux's favor. For example, this cluster is 277 nodes with better networking, and we chose Linux over Tru64, due to Linux's super system administration capabilities.
BTW, you can get Compaq's great Alpha compilers for Linux.
Almost no one uses Linda -- what would you think UDel does?
Most people with systems like this use a batch queue system like PBS and message passing libraries like MPI.
This source of funding isn't that unusual -- the University of Virginia Centurion cluster was funded by two $450,000 MRI grants.
Prior art?
Try the "Hydra" operating system, late 1960's.
Just about any form of distributed computing is prior art on this one.
You don't get it. Omega 1 means the universe expands forever.
I find it funny that no one challenged Tom's claim that Unisys invested resources to invent LZW.
LZW was invented by two guys in Israel, Lempel and Ziv. Unisys (and Welch) patented their work, and doesn't give them a dime. It's completely legal, yet stupid. This case has always been the poster child of the "Patents are Lame" movement.
Whenever you see Unisys whining about all the money they spent to develop LZW, remember the facts.
See: http://www.faqs.org/f aqs/compression-faq/part1/section-7.html.