the McDonalds in northern Maine sell Poutine. Lots of people also have French-Canadian accents. I live in Bangor, ME which is considered 'Northern Maine' as far as population density, but geographically would be
slightly south of the center of Maine's landmass. If I drive the 3-4 hours to the northern tip of Maine( i recently had to go to a college friends wedding up there), it's like a whole different country.
I'd like to see an iCube: singe G5 processor, with a style similar to the old G4 cubes, but aluminum to match the G5 powermacs. That way people can get an entry level G5 without having to replace their monitor the same time they purchase their new single G5.
BTW, for some off topic Apple news, I just started recieving 259 Xserve raids at work... fun fun! http://www.clusters.umaine.edu/gallery/xserv es
without patching OpenPBS does NOT scale well. Altair Engineering says it doesn't scale well over 32 nodes. They short-sell it a little there to try to sell PBS Pro, but I know from experience it is VERY flakey with no patches.
OpenPBS is shit. It needs about 25 3rd party patches(many of which have to be applied in order) to be halfway decent, (Altair Engineering doesn't actively develop it anymore so these patches won't be integrated into OpenPBS). Also it often falls flat on its face. I wouldn't want to use OpenPBS unless I had a trivially small cluster.
If you want something free, TORQUE is OK. It is a OpenPBS derivative (they started with the last OpenPBS version and added all the popular scalability and fault tolerance patches). TORQUE is actively developed under some DOE contracts, and even the company that has the DOE contracts to develop and support TORQUE will give other people some free support (I've had them on the phone helping to debug some of the code). You can get TORQUE from Supercluster.org.
PBS Pro is very good, but costs a lot unless you are a degree granting department (then it is free).
if you live in such a violent (from an EU perspective) country as USA.
Someone in London is more likely to get mugged, beated, or stabbed than someone in NYC...
Where I'm from I can leave my door unlocked while I'm out with little chance of anyone stealing something from me.
Also are you saing the USA is more violent than some incoming EU members? Turkey, Cypress, the Czech, Republic, Latvia, Estonia, Slovakia? Maybe. I don't know, but I don't know if I believe that
Usually people in the supercomputing community don't call loosely coupled groups of computers on the internet running some distributed problem solver a cluster. Usually cluster refers to a more tightly coupled system.
The beauty of programming is you don't have to be located in LA to make it as a tech company. Some software companies should move to where I live - Maine. You can buy a very large, very nice 5 br house in the Bangor Maine area for less than what you paid for your house. This is near the University of Maine, which has a good college of engineering(we offer more acredited BS engineering degrees than any other university in New England), and a small, but good, computer science department with a phd program. We have new engineering facilites going up all over campus. All these graduates are leaving the state to find work. Companies can pay them less to stay in Maine, and their standard of living will still be the same.
Shuttles were designed to fly much more often than they do now. NASA had envisioned at least bi-weekly flights when they designed the shuttle. At that rate of launch there actually would have been a cost savings over an entirely disposable system.
They have not been overworked. They were built to fly at least 100 missions without major overhauls. Columbia had completed 38 missions before the disaster.
Now this was supposed to be in a much shorter timeframe, but its the number of missions, not age, that causes stress on the shuttle. Also they had just done an overhaul of the Columbia before the disaster, so they did shorten the number of missions between overhauls.
I've read recent articles that NASA plans on keeping the remaining three shuttles flying for another 20 years. They plan on doing this with smaller crews, using the shuttle to tote cargo, and speeding up development of the space plane to bring crews back and forth to the space station. The reduced crew of the shuttle would make an ejection seat a viable option.
I thought it might be a common name, but figured I'd give it a shot. I have some fond memories of drinking beer and playing gamecube with Rahul after our "Computers, Ethics, and Society" class.
I don't think I've touched a gamecube since then.
I think he had moved back to Tennessee(where his parents live) after he graduated(we went to a New England University - where I'm now a graduate student)
Since this is about Maine, and I live in Maine I thought I'd add a little about Maine laws:
Actually in Maine the age of consent is 16, plus there is a minnimum age difference(4 years) so if a 15 year old has sex with a 18 year old it will not break the law, since the age difference is less than 4 years.
None of this applies if the older person is a teacher. If a teacher has sex with a student they can be charged with statutory rape even if the age gap is less than 4 years or the student is over 16.
The Altix isn't a 256 CPU single image system, at the current time anyway.
Altix nodes come with up to 64 Itanium2 processors. Each node has a single OS image. A 256 CPU Altix supercomputer would be made up of at least 4 nodes(you could use 8 32 CPU nodes,etc). Each of these nodes has its own OS. There is shared memory accross nodes however, using SGI NUMALink between nodes.
I looked into the Altix when I thought we might want a 64 processor Itanium system, but our main code isn't network bound, so there was no way to justify the extra cost of the Altix.
Lamport, Leslie. "Time, Clocks, and the Ordering of Events in a Distributed System", pps. 558-565, CACM, Vol. 21, No. 7, July, 1978
Parnas, David. "On the Criteria To Be Used in Decomposing Systems into Modules", pps. 1053-1058, CACM, Vol. 15, No. 12, December, 1972
Hester, S.D, Parnas, D.L., and Utter, D.F. "Using Documentation as a Software Design Medium", pps. 1942-1977, The Bell System Technical Journal, October, 1981
Parnas, David L. and Clements, Paul C. "A Rational Design Process: How and Why to Fake It", Presented at the Tapsoft Joint Conference on Theory and Practice of Software Development, Berlin, March, 1985
And one more(I know this guy, and I had to read this for a class, so I figured I'd give him some props..)
Wheeler, Tom. Software System Development Through The Use of Formal Documentation, Ch. 2 (System Documentation), PhD. Thesis, Steven's Institute of Technology 1988
This SGI isn't a beowulf cluster. Traditionally beowulf clusters refer to clusters that use COTS hardware, don't have global shared memory, etc. Lots of people in the cluster community won't even call clusters of workstations beowulf clusters if they have some high speed network like Myrinet. We just call ours a Linux cluster, a cluster, a distributed memory supercomputer...
You can program your beowulf cluster in C or Fortran using a free MPI(message passing interface) implementation called MPICH. I have even seen a scaled down version of MPI for Python, (which requires MPICH to use). So start learning MPI. MPI-1 has 129 functions, but you can write most programs using a small subset of these calls.
If you don't want to pay much money I suggest using C, because g77 sucks and there are no free Fortran 90 compilers. We use the Portland Group Fortran and C compilers as well as the Intel Fortran Compiler. I think we are going to switch completely to Intel Fortran and C.
Why do you want to use a beowulf cluster if you have no clue about them or parallel programming in general? Just because they are 'cool'? A beowulf cluster is very usefull for modeling or datamining, but unless you are running models that take days/weeks/months on your workstation you won't need the processing power of a cluster. Right now we have someone running a model on 76 processors that takes about 9 hours to finish a 1 year cycle in the model. They want to run the model for a total of 50 years. This is a model of the pacific ocean where they introduce carbon into the ocean, and then they see what effect that has on temperature change, etc. After they get their 50 year resluts for the Pacific they want to do a global simulation. This is the real use of beowulf clusters. They aren't for load ballancing web servers, playing quake, or any of the other things people post about every time there is an article about supercomputers/beowulf clusters.
The speed up you will get really depends on your application. The more communication is necessary, the smaller the speed up will be. If you have a 5 node cluster, with 2 processors per node, the theoretical maximum speed-up is 10, but you will never achieve that because of parallel overhead(MPI calls, communication time, etc).
If you want more information on parallel programming and cluster computing send me a private message telling me what you hope to do with your cluster.
Fortran is FAR from dead. It is used extensively in high performace computing. Computational chemistry, biology, CFD,... most of this is still Fortran.
See what simulations people run on top500.org computers. Most of that code will be written in Fortan.
Slashdot Mirror
the McDonalds in northern Maine sell Poutine. Lots of people also have French-Canadian accents. I live in Bangor, ME which is considered 'Northern Maine' as far as population density, but geographically would be slightly south of the center of Maine's landmass. If I drive the 3-4 hours to the northern tip of Maine( i recently had to go to a college friends wedding up there), it's like a whole different country.
Opps I didn't mean Xserve RAIDS - I don't know why I was thinking about RAID.... 259 Xerves - 1 Xserve RAID
I'd like to see an iCube: singe G5 processor, with a style similar to the old G4 cubes, but aluminum to match the G5 powermacs. That way people can get an entry level G5 without having to replace their monitor the same time they purchase their new single G5.
v es
BTW, for some off topic Apple news, I just started recieving 259 Xserve raids at work... fun fun!
http://www.clusters.umaine.edu/gallery/xser
You are correct. OS X is based on BSD/MACH, but they didn't get official UNIX(TM) classification because it costs $$ or something like that...
without patching OpenPBS does NOT scale well. Altair Engineering says it doesn't scale well over 32 nodes. They short-sell it a little there to try to sell PBS Pro, but I know from experience it is VERY flakey with no patches.
OpenPBS is shit. It needs about 25 3rd party patches(many of which have to be applied in order) to be halfway decent, (Altair Engineering doesn't actively develop it anymore so these patches won't be integrated into OpenPBS). Also it often falls flat on its face. I wouldn't want to use OpenPBS unless I had a trivially small cluster.
If you want something free, TORQUE is OK. It is a OpenPBS derivative (they started with the last OpenPBS version and added all the popular scalability and fault tolerance patches). TORQUE is actively developed under some DOE contracts, and even the company that has the DOE contracts to develop and support TORQUE will give other people some free support (I've had them on the phone helping to debug some of the code). You can get TORQUE from Supercluster.org.
PBS Pro is very good, but costs a lot unless you are a degree granting department (then it is free).
Usually people in the supercomputing community don't call loosely coupled groups of computers on the internet running some distributed problem solver a cluster. Usually cluster refers to a more tightly coupled system.
The beauty of programming is you don't have to be located in LA to make it as a tech company. Some software companies should move to where I live - Maine. You can buy a very large, very nice 5 br house in the Bangor Maine area for less than what you paid for your house. This is near the University of Maine, which has a good college of engineering(we offer more acredited BS engineering degrees than any other university in New England), and a small, but good, computer science department with a phd program. We have new engineering facilites going up all over campus. All these graduates are leaving the state to find work. Companies can pay them less to stay in Maine, and their standard of living will still be the same.
They have not been overworked. They were built to fly at least 100 missions without major overhauls. Columbia had completed 38 missions before the disaster.
Now this was supposed to be in a much shorter timeframe, but its the number of missions, not age, that causes stress on the shuttle. Also they had just done an overhaul of the Columbia before the disaster, so they did shorten the number of missions between overhauls.
I've read recent articles that NASA plans on keeping the remaining three shuttles flying for another 20 years. They plan on doing this with smaller crews, using the shuttle to tote cargo, and speeding up development of the space plane to bring crews back and forth to the space station. The reduced crew of the shuttle would make an ejection seat a viable option.
I think he had moved back to Tennessee(where his parents live) after he graduated(we went to a New England University - where I'm now a graduate student)
Hey, I went to college with a Rahul... Do you work for Ebay?
Actually in Maine the age of consent is 16, plus there is a minnimum age difference(4 years) so if a 15 year old has sex with a 18 year old it will not break the law, since the age difference is less than 4 years.
None of this applies if the older person is a teacher. If a teacher has sex with a student they can be charged with statutory rape even if the age gap is less than 4 years or the student is over 16.
Altix nodes come with up to 64 Itanium2 processors. Each node has a single OS image. A 256 CPU Altix supercomputer would be made up of at least 4 nodes(you could use 8 32 CPU nodes,etc). Each of these nodes has its own OS. There is shared memory accross nodes however, using SGI NUMALink between nodes.
I looked into the Altix when I thought we might want a 64 processor Itanium system, but our main code isn't network bound, so there was no way to justify the extra cost of the Altix.
Lamport, Leslie. "Time, Clocks, and the Ordering of Events in a Distributed System", pps. 558-565, CACM, Vol. 21, No. 7, July, 1978
Parnas, David. "On the Criteria To Be Used in Decomposing Systems into Modules", pps. 1053-1058, CACM, Vol. 15, No. 12, December, 1972
Hester, S.D, Parnas, D.L., and Utter, D.F. "Using Documentation as a Software Design Medium", pps. 1942-1977, The Bell System Technical Journal, October, 1981
Parnas, David L. and Clements, Paul C. "A Rational Design Process: How and Why to Fake It", Presented at the Tapsoft Joint Conference on Theory and Practice of Software Development, Berlin, March, 1985
And one more(I know this guy, and I had to read this for a class, so I figured I'd give him some props..) Wheeler, Tom. Software System Development Through The Use of Formal Documentation, Ch. 2 (System Documentation), PhD. Thesis, Steven's Institute of Technology 1988
This SGI isn't a beowulf cluster. Traditionally beowulf clusters refer to clusters that use COTS hardware, don't have global shared memory, etc. Lots of people in the cluster community won't even call clusters of workstations beowulf clusters if they have some high speed network like Myrinet. We just call ours a Linux cluster, a cluster, a distributed memory supercomputer... You can program your beowulf cluster in C or Fortran using a free MPI(message passing interface) implementation called MPICH. I have even seen a scaled down version of MPI for Python, (which requires MPICH to use). So start learning MPI. MPI-1 has 129 functions, but you can write most programs using a small subset of these calls. If you don't want to pay much money I suggest using C, because g77 sucks and there are no free Fortran 90 compilers. We use the Portland Group Fortran and C compilers as well as the Intel Fortran Compiler. I think we are going to switch completely to Intel Fortran and C. Why do you want to use a beowulf cluster if you have no clue about them or parallel programming in general? Just because they are 'cool'? A beowulf cluster is very usefull for modeling or datamining, but unless you are running models that take days/weeks/months on your workstation you won't need the processing power of a cluster. Right now we have someone running a model on 76 processors that takes about 9 hours to finish a 1 year cycle in the model. They want to run the model for a total of 50 years. This is a model of the pacific ocean where they introduce carbon into the ocean, and then they see what effect that has on temperature change, etc. After they get their 50 year resluts for the Pacific they want to do a global simulation. This is the real use of beowulf clusters. They aren't for load ballancing web servers, playing quake, or any of the other things people post about every time there is an article about supercomputers/beowulf clusters. The speed up you will get really depends on your application. The more communication is necessary, the smaller the speed up will be. If you have a 5 node cluster, with 2 processors per node, the theoretical maximum speed-up is 10, but you will never achieve that because of parallel overhead(MPI calls, communication time, etc). If you want more information on parallel programming and cluster computing send me a private message telling me what you hope to do with your cluster.
Fortran is FAR from dead. It is used extensively in high performace computing. Computational chemistry, biology, CFD, ... most of this is still Fortran.
See what simulations people run on top500.org computers. Most of that code will be written in Fortan.