You know how sometimes you read something and your heart starts racing and your tail starts wagging because you know *exactly* what the writer is talking about? This is it!
So, what about us poor grad students? I went to 3 prospective advisers at the beginning of last year: one's doing augmented reality, the other one bioinfo, the last ODEs. All of them had defense or army funding, and I had a problem with that. So I brought up as nicely as I could the fact that I objected to doing research that was being funded by military institutions.
The first one suggested that I figure out my priorities early on in my academic career and I would always be a loser if I chose not to take the money where I could get it. The second said it would be better if peace-loving folk like "us" took the money and did something with it that had peaceful applications. The third ignored me.
Frankly, I'm not satisfied with any of those attitudes. I suppose I could rant on and on about why, but the point is, grad students have to live with this climate and we don't always like it. Maybe that's even when the problem starts... You don't have a choice at the beginning of your career, when you're actually thinking intensely about this stuff, and maybe you get stuck feeling like you have no choice later, when you really do, and when you could be giving your underlings a choice too.
I know you mentioned you already have access to a beowulf type cluster ("like Los Alamos", but they're really *everywhere* now, and LANL is not even doing the most interesting work on them), but that your CFD code won't run on linux.
This needs a little more looking into: why won't it run on linux? I assume it's Fortran (or C, in which case you're super lucky). What code compiles on IRIX and HP-UX, of all things, but doesn't compile under linux? Have you tried proprietary (eg PGI) compilers? Is this a library issue? If I were you I'd look into that angle a little more.
Why would I pick a beowulf?
- transparency: no proprietary clustering, SMP or other code. You will need to use either home-coked or open source tools to administer it (ANL's chiba tools being an example of the type), but that is much more transparent than either HP's or IBM's ways.
- Linux: if you have a problem, you have a huge base of users and admins to talk to (ExtremeLinux folks, Beowulf underground folks etc etc). Open source makes problems ultimately the admin's responsibility, and you might not like that accountability, of course. But if you've dealt with large company support, you'd know it's not the best thing to depend on.
- Cost: reasonable beowulfing hardware costs a small fraction of what O2k's, SP's or that class of machine cost.
- Reusability/Extensibility: PCs are easy to downgrade to desktops after a while, just by adding a good video and sound card.
- Scalability: first off, I know from personally having administered beowulfs larger than 24 boxes, that there are next to no scalability issues with that order of number of machines. On the other hand, you can add on as you go. Beowulfs typically perform well in inhomogeneous environments.
I'd go with Alphas, since CFD + a lot of other scientific computing applications in physics and chem are fp intensive.
Hope this is somewhat comprehensive.
Btw, everyone's doing this: it's not risky anymore. On the other hand the beowulf approach has its merits, it's not a fad. That is to say, I think linux clusters, as opposed to SPs are here to stay.
This has no scientific merit whatsoever, but it's monitors that do it to me. It takes a long long long time for me to get to a nasty block, but when I do, it usually takes a nap and a switch to a laptop to fix it. Somehow, it's the groundedness of being in front of a desktop monitor (+ the radiation, but how to explain that?) that keeps me from thinking outside the frame. I usually just need to break the frame of my solution to whatever programming problem I'm trying to solve, and do something unorthodox, and the first step is always to feel mobile and unrestricted. Switch to the laptop after a nap, hell! take the laptop to starbucks or a bar or the beach... Then I feel hardcore. Take this as 2c worth, like it is.
The thought has occured to college students all over the US. I took myself to task one day, and attended an information session, designed to answer exactly the question "What does it take to work for the NSA?" One interesting thing to note was that the talk was delivered by a female mathematician, who was about to leave the NSA to start work in the private sector. I wasn't sure what to make of that... Interesting little tidbits that arose from my conversation with this soon-to-be-former employee of the NSA: * unlike a University, there are no non-US citizens working for the NSA. I presume that makes it slightly boring. I've found the international students and faculty in my U. to be probably the most informed and amusing to hand out with. * the publish-or-perish issue is a non-issue. You cannot publish new findings until they're not new. Personally, I'd have a hard time spooning that one down. * You do get to work with smart people. I don't know what your definition of genius is, so I won't go there. I should hope (not for the sake of National Security or any of the Nationalist reasons that are often given, but simply because I really *want* to believe we're not governed by idiots) that the NSA chooses employees better that the rest of the *.gov. * they say they try to cooperate with business. I don't know how they go about it though.
In any case, this is getting long winded, but I think it summarizes my thoughts pretty well. 1) you have to sizzle before you publish, 2) it might get a little single-minded at times.
Slashdot Mirror
You know how sometimes you read something and your heart starts racing and your tail starts wagging because you know *exactly* what the writer is talking about? This is it!
So, what about us poor grad students? I went to 3 prospective advisers at the beginning of last year: one's doing augmented reality, the other one bioinfo, the last ODEs. All of them had defense or army funding, and I had a problem with that. So I brought up as nicely as I could the fact that I objected to doing research that was being funded by military institutions.
The first one suggested that I figure out my priorities early on in my academic career and I would always be a loser if I chose not to take the money where I could get it. The second said it would be better if peace-loving folk like "us" took the money and did something with it that had peaceful applications. The third ignored me.
Frankly, I'm not satisfied with any of those attitudes. I suppose I could rant on and on about why, but the point is, grad students have to live with this climate and we don't always like it. Maybe that's even when the problem starts... You don't have a choice at the beginning of your career, when you're actually thinking intensely about this stuff, and maybe you get stuck feeling like you have no choice later, when you really do, and when you could be giving your underlings a choice too.
Sorry, this became a rant...
I know you mentioned you already have access to a beowulf type cluster ("like Los Alamos", but they're really *everywhere* now, and LANL is not even doing the most interesting work on them), but that your CFD code won't run on linux.
This needs a little more looking into: why won't it run on linux? I assume it's Fortran (or C, in which case you're super lucky). What code compiles on IRIX and HP-UX, of all things, but doesn't compile under linux? Have you tried proprietary (eg PGI) compilers? Is this a library issue? If I were you I'd look into that angle a little more.
Why would I pick a beowulf?
- transparency: no proprietary clustering, SMP or other code. You will need to use either home-coked or open source tools to administer it (ANL's chiba tools being an example of the type), but that is much more transparent than either HP's or IBM's ways.
- Linux: if you have a problem, you have a huge base of users and admins to talk to (ExtremeLinux folks, Beowulf underground folks etc etc). Open source makes problems ultimately the admin's responsibility, and you might not like that accountability, of course. But if you've dealt with large company support, you'd know it's not the best thing to depend on.
- Cost: reasonable beowulfing hardware costs a small fraction of what O2k's, SP's or that class of machine cost.
- Reusability/Extensibility: PCs are easy to downgrade to desktops after a while, just by adding a good video and sound card.
- Scalability: first off, I know from personally having administered beowulfs larger than 24 boxes, that there are next to no scalability issues with that order of number of machines. On the other hand, you can add on as you go. Beowulfs typically perform well in inhomogeneous environments.
I'd go with Alphas, since CFD + a lot of other scientific computing applications in physics and chem are fp intensive.
Hope this is somewhat comprehensive.
Btw, everyone's doing this: it's not risky anymore. On the other hand the beowulf approach has its merits, it's not a fad. That is to say, I think linux clusters, as opposed to SPs are here to stay.
--dubido
This has no scientific merit whatsoever, but it's monitors that do it to me. It takes a long long long time for me to get to a nasty block, but when I do, it usually takes a nap and a switch to a laptop to fix it. Somehow, it's the groundedness of being in front of a desktop monitor (+ the radiation, but how to explain that?) that keeps me from thinking outside the frame.
I usually just need to break the frame of my solution to whatever programming problem I'm trying to solve, and do something unorthodox, and the first step is always to feel mobile and unrestricted. Switch to the laptop after a nap, hell! take the laptop to starbucks or a bar or the beach... Then I feel hardcore.
Take this as 2c worth, like it is.
The thought has occured to college students all over the US. I took myself to task one day, and attended an information session, designed to answer exactly the question "What does it take to work for the NSA?"
One interesting thing to note was that the talk was delivered by a female mathematician, who was about to leave the NSA to start work in the private sector. I wasn't sure what to make of that...
Interesting little tidbits that arose from my conversation with this soon-to-be-former employee of the NSA:
* unlike a University, there are no non-US citizens working for the NSA. I presume that makes it slightly boring. I've found the international students and faculty in my U. to be probably the most informed and amusing to hand out with.
* the publish-or-perish issue is a non-issue. You cannot publish new findings until they're not new. Personally, I'd have a hard time spooning that one down.
* You do get to work with smart people. I don't know what your definition of genius is, so I won't go there. I should hope (not for the sake of National Security or any of the Nationalist reasons that are often given, but simply because I really *want* to believe we're not governed by idiots) that the NSA chooses employees better that the rest of the *.gov.
* they say they try to cooperate with business. I don't know how they go about it though.
In any case, this is getting long winded, but I think it summarizes my thoughts pretty well. 1) you have to sizzle before you publish, 2) it might get a little single-minded at times.
--TL