Daniel Kennefick interview with Andrew Abrahams (contd.)

0:00 Change that allows message passing between different processors. So you can have, say you split up your computational domain into, if you have 30 processors, into 30 different patches and essentially what happens to each patch evolves and then it coordinates its boundaries with all the other patches and then so they all have correct information that they can use. What he did was really monumental piece of work and apparently has been used by people in other fields as well and it's starting to be used in ours. This empire code, which is the one based on the hyperbolic formulation, I mentioned that it was discontinued, was actually built entirely based on that infrastructure. So DAG stands for Dynamic Adaptive Grid Hierarchy, in case you want it. Now we're basically in a situation where the other code, the ADM code, is now using DAG. But there have been various bugs and things, and it's taking some time. So, that's mainly. Now, this guy, Manish Parishar, who developed it, then went to Texas and worked with the other computer scientists there, and he basically continued developing that while he was there. He spends time in Potsdam also and is a part of Ed's group.

2:30 What's his name? Manish is his first name. His last name is Parashar. P-A-R-A-S-H-A-R. It may be an I actually, P-A-R-A-S-H-A-R. If you look on the website, you might find out. So, anyway, so the impact has continued to be fairly active. The two people who have been here were working, collaborating with Ed's group on various projects, none of which really, I think, had any big impact on the alliance. So, it was certainly very important politically to have computer science. There has been some, but it hasn't been quite lived up to as well as we hoped. So, there were at least these two different codes or projects based on the talk about it? Yeah. And then there's a whole set of codes that come from, that came from the CSA, which had sort of, yeah, they're not involved in the Maya Center directly anymore, but they had a code based on a different hyperbolic form of validation, which was written by John Massello, and then that code was called the H-code, I think. Then, they also had sort of their historic, they had a code called the G-code, which was their more standard ADM style code.

5:00 So they also, they actually worked with both, the ADM and the mathematical code? Right. I mean, they were different people, but, you know, we were in the same group. The microbiology approach, at least temporarily, not based on any real scientific merit issues, so much that it's more efficient for some. So they've continued with both. I'm sure Ed will give you a much more complete version of all the different codes. It'll be a lot. So I guess the size of the groups vary a lot. The NCSA group seems to... It used to be very big. Or what used to be. Yeah. Well, yeah, they always had, well, actually the genesis of it is that David Hobel came here sort of in like 1988, I guess right before I left. And that's where his postdoc started building the group. And then Ed came as a postdoc probably in 89 or so. Then eventually David left and Ed continued building through so they have they always have at least one postdoc I think and several graduates so it was doesn't sound big from other fields but actually was one of the larger efforts so they always had probably two postdocs and a set of graduates Cornell had two faculty and typically two or three postdocs and a year graduate student. The sizes are not that different. I think it's more of a matter of focus on different things.

7:30 Yeah, other areas that they also worked on were things that weren't so obviously focused. Also, Penn State, right, so probably one of the real leaders of the alliance, at least as far as getting people to work together and bringing the various political gaps that have come up. Also, they're having a pretty wide-ranging and also fair understanding of the science. So, he's always had postdocs in there, or I guess probably one postdoc that I've seen. And then Sam Finn is also, you know, sort of loosely connected to members of the Alliance. But he hasn't been so susceptible to anybody? Well, I mean, of course he's very involved in a lot of good stuff, and so he's been probably, I think, heavily. Okay, so there was another, you know, very early on, or pretty early on, there was a decision made not to, to, there was a decision made to focus on finite difference in methods for solving equations. Sam was very interested in doing something with spectral techniques. Once that sort of was pushed aside, then, you know, we also wanted to compare, we wanted to compare, so we didn't want to do anything else. Yeah, yeah, yeah, that's interesting. Do you have licenses for him? Yeah, I have a problem with him. I think he was also rather fed up with all the politics of it all.

10:00 Yeah, I get the exact same impression. Not that I had spoken to him. I thought he was something formally quite wild. Yeah, I'll probably catch you. Long enough then. That's a lot of useful information for me. Thanks very much. Well, I think that once you're just going to check, ask you who else around here is still in love, I guess I'm going to talk to Peter, I mean, and the people in the institution here, who else around here is still actively in love? I think that's probably, you know, a good question. I mean, one person you might try to talk to is Lucasius. I think one of these geosciences was involved in, sort of more early on in the Amayas, and now is working on the neutron star, is involved in the neutron star, and Paul Scaler, and some kind of other people. I'll try that too, because I'd like to talk to the neutron star people as well, and it's hard to fit everything in. Yeah, so, I mean, I expect he's not terribly intact. Well, that should be an interesting topic for me too, but now you want to read less from it. Well, thank you very much.