00:01:28.199 - 00:01:49.650

C 5 H 5 102. C 5 H5N2. I resist on it. Right. What am I going to do about the final I want? I don't know. Did you want the fire? I don't know. I might have. It's hard to say. How we get in. A linear

00:01:49.650 - 00:02:11.369

systems and communications and control. 401. Yes sir. I should have aced it. You should ice it, yeah, I know. See if there's a wire in C4M3 pin 2. Right. Yes there is. Well, good. How about satering it? What's the sound? Well, I've got it on this

00:02:11.369 - 00:02:38.270

side. You're going to need it, I guess. It Injury into We call it the R2 for sure. The computer will be, well, it'll do the things that That you'd expect that most computers would do, but in addition to that, uh, we hope to try some different

00:02:38.279 - 00:02:55.229

applications of computers, some applications involving graphics, constructing pictures with the machine and solving some problems graphically, some problems in geology, for instance. We're right now in the process of completing the construction of the machine. The frame I'm working on is the control frame of the computer.

00:02:55.479 - 00:03:14.970

It decodes instructions. Passes along commands to other elements of the machine and tells them what to do. The machine's unusual in that. Not only was it designed here on Rice, but it's being built by technicians and undergraduates and Of course some of the faculty Provides a

00:03:14.970 - 00:03:37.179

really unusual opportunity for students to get hands-on experience in in computer science. The question is, when you get ready To compile the code or you're going to know where this thing wants to be and the answer is no unless you go back and patch it up

00:03:37.179 - 00:03:48.949

because you haven't set up any of this stuff, OK? You're you're you're working your way through this tree and you don't have any idea. How much code is going to be turned out underneath the tree, and the code that's turned out underneath the tree corresponds exactly

00:03:48.949 - 00:04:02.529

to the code for these things, OK? So you surely can't do it unless you sort of back up and do it, but I think you can back up and do it. So we're talking still about two different issues. One is how to cope with the problem

00:04:02.529 - 00:04:20.309

of this tree, and the other is what, what sort of code do you turn out, and I think. I think that you want to generate code when you enter the on block entry to copy the display, OK, copy the display. doing Copy the old display into

00:04:20.309 - 00:04:31.609

here, OK, OK, OK, copy, copy that pointer right there or if you're here copy those pointers there. You want to set up the stack top, OK, using the fixed date area size and then just go ahead and proceed and turn out the seat node, turn out

00:04:31.609 - 00:04:43.070

the code for the seat node once you've done those things. I'm not sure. I'm, I'm not sure of the answer to the problem. It could be possible that with blocks, since blocks are fixed, that that can all be done at compile time. But you cannot, you

00:04:43.070 - 00:04:59.230

surely cannot do it. No, it can't be done to compile time for this reason, well, Yeah, it can be done at compiled time because I was, I was thinking the reason that you couldn't is because suppose you have a procedure in here, OK? You, you never

00:04:59.230 - 00:05:26.239

know when the procedure is going to be called. It might be called and it might not be called. I don't know Seems to me, maybe you've got to do it at runtime. The whole point of doing dynamic storage allocation, OK, is to save space with stuff

00:05:26.239 - 00:05:36.600

that's going to be temporary or maybe variable in length, and I think what the stack top says is, OK, we're executing in level 0. We start executing in level 1, we start executing in level 2, we execute in level 3, we leave level 3, maybe we

00:05:36.600 - 00:05:48.429

get all the way back up to level 1. That says then the fact that we've set up each of these stack tops and they're known at compile time. Well, they're almost, they're not really known at compile time. The places where they're going to be is known

00:05:48.429 - 00:06:00.950

at compile time, OK? They're sort of seated or set up at runtime. The answer is though, when you get to here, you know that all this area above the orange line is, is OK now to use for dynamic storage allocation. When you exit block 3, you

00:06:00.950 - 00:07:38.200

no longer going to refer to those variables. It's outside their scope or their range, OK. Code generation isn't easy. It certainly isn't. Oh Thanks. I go back and get on the sho. You And OK. Yeah. So I had no preconceptions and and just nothing and

00:07:38.200 - 00:07:53.040

it just came and happened and I I liked it and the people are really nice they're friendly. I like the way the classes are set up. It's kind of a relaxed atmosphere and um they have more opportunities to get out and do things, but the problem

00:07:53.040 - 00:08:13.700

is that you do them on your own. You don't do them. Nobody's pushing you into anything, you know, and it's like you can go out and do exactly what you want. If you don't want to do it, you don't do it. I, I don't know, I

00:08:13.700 - 00:08:23.700

spend a lot of time at the tennis courts because, you know, you're sitting around and you're studying all the time and you just get kind of bored. Well, not bored, you just after a while you just can't soak them anymore and you have to go out

00:08:23.700 - 00:08:38.969

and do something and I usually go out to the courts and and play tennis because you're out there and you're running around physically and then also your mind's just in another world, you know, and, uh. I'm not really a competitive tennis player. It's all just kind

00:08:38.969 - 00:08:54.369

of for the, it's relaxation for me, you know, she smiles. I'm out there all the time you're not competitive. You just wanna play. Do you play tennis? Let's let's have a match. Oh, no, no. Um, yeah, this past Sunday we went, we played 6 sets, and

00:08:54.369 - 00:09:08.719

it just happened to be, you know, the Saturday the cold front came in, it's 58 degrees. We're going, wow, wow, Nene running through the quad just singing and dancing. Nobody was out there o'clock in the morning. The sun's coming up and it's it's kind of misty

00:09:08.719 - 00:09:26.169

and cold and everything. It's really neat. We should do that more often, but it won't, mhm, but uh. We're waiting for the next cold front to come in. Yeah, we've been watching the weather map. Oh, we've got a low. Oh wow, weather because I'm, I'm really

00:09:26.169 - 00:09:52.840

getting tired of the humidity and stuff because it kind of brings you down because you can't do everything. In the rain too. That's Houston. This is really dumb. Let's talk about something else. I don't know. Think of something. How about engineering. How about it. Well What

00:09:52.840 - 00:10:09.640

do you want to do? I mean, what kind of engineering biomed. Um, well, it's really strange. I was under the impression that Rice had an undergraduate program here and so I came up all enthusiastic about it. Come to find out they have a graduate program there

00:10:09.640 - 00:10:22.280

but they don't have an undergraduate, so I have to go through. The what you get an undergraduate panel or something after they decide your courses, what you need for a degree so I'm gonna have to go talk to them. But in the meanwhile, you know, you

00:10:22.280 - 00:10:39.369

just take the regular engineering courses, your, your freshman engineering courses, math, you was doing physics, that's about all you can handle and uh you know what is it? Is this what you mean? Like you work in a hospital and um you choose. OK, like a doctor

00:10:39.369 - 00:10:52.729

is a doctor and he really doesn't know everything about the equipment he's using and so I need somebody that's, that's got the knowledge and like the engineer, this is a new field with all the new equipment coming up for doctors use in the operating room and

00:10:52.729 - 00:11:08.419

everything and um. He's half doctor, half engineer. He chooses the equipment to use during the operation and he also, you know, maintains and stuff. One of the really interesting things is like he monitors the, the premature babies. They, they design these little things where you just

00:11:08.419 - 00:11:21.460

stick them on the baby and monitor their heartbeat and their pulse and their temperature and and all this type of stuff and you sit there and you can play it all the ms, but I mean it's, it's different from engineering in the respect that. You're not

00:11:21.460 - 00:11:36.239

sitting at a desk and you're not, you know, working with formula. Yeah, you've got people for once, and this is really kind of important to me because I I don't know. People, I, I like people and I like to get to know them and stuff, and

00:11:36.239 - 00:11:52.500

when you're helping people, you know, it, it becomes a different story altogether than just sitting at a desk when you're helping somebody like building them a road or or doing something like that, so um. It's really it's it's something it's a new field and it's kind

00:11:52.500 - 00:12:08.179

of exciting and if you're really ambitious you get your PhD in it and and you can start designing stuff like you know for the the artificial hearts and stuff like this and the pacemakers and, and, uh, I don't know it just seems like something new and

00:12:08.179 - 00:12:40.450

exciting and yet you can meet people. I guess. I can't see me, you know, designing bridges and building roads or, you know, doing stuff like that. My name's Tom Dye, and I graduated from Rice a year ago and I'm working now here in Albuquerque with the

00:12:40.450 - 00:13:07.750

Bureau of Sport Fisheries and Wildlife. It's a federal organization and we have a responsibility for wildlife refuges and national fish hatcheries in the 8 state region. And my job is to. be concerned with all the pollution aspects of our activities in the field stations. A year

00:13:07.750 - 00:13:35.059

ago we planted 2,900,000 plus fish weighing a total of 136,000 pounds. This past year, we had about 106,000 pounds production. And our numbers ran 2,400,000 last year we had a monitoring program here in this region where we Kept track of various parameters over a 12 month

00:13:35.059 - 00:13:58.950

period at 4 of our cold water hatcheries and from this data and the data that we found in the literature, uh, we were able to get a good idea of what exactly fish hatchery effluent contained and after this. We could get an idea of how much

00:13:58.950 - 00:14:20.020

treatment would be necessary to bring it down to a level where we thought it would not cause any trouble in the streams that we would be discharging to. For each of the parameters we're concerned with, such as biological oxygen demand, suspended solids, phosphorus compounds, nitrogen compounds,

00:14:20.030 - 00:14:38.104

we specified a certain number of pounds of these pollutants that we would allow in the effluent per pound. Of fish food that we added to our system. So if you know the amount of food you used in any one day, you can just use the conversion

00:14:38.104 - 00:15:01.219

factor, which is the standard, and come up with the maximum level of pollutants that will be allowed in the effluent pipe. And another thing about the standards that's important. We believe that the direction of standard setting in this country will be Toward effluent standards, that is

00:15:01.219 - 00:15:22.099

the concentration or the amount of pounds in the effluent pipe rather than stream standards, which are usually the current way of looking at things and that specifies a certain mixing zone, but in our standards and probably future standards for other industries will be in terms of

00:15:22.099 - 00:16:07.950

just no mixing, but rather the uh. Amount of pollutants actually coming out of the pipe. I don't get one. Yeah OK. Well, of course is very powerful. With their lobbies and their money in Washington. If they accept money for their campaign or something from a

00:16:07.950 - 00:16:34.570

big company and then. They get flack if they are strong on pollution control. But if we don't clean up. There's no use having a good job if you can't breathe the air. I drank the water. That's the, that's the whole, that's the thing that's really bad

00:16:34.570 - 00:16:48.549

about my cogeneration scheme. I haven't organized the difference between modes like my tends thesis, references and values the same problem hit me when I did. Uh, the evaluation of conditionals in Sal 68. OK, but all I'm trying to point out is that that you may need

00:16:48.549 - 00:17:01.400

in your symbol table an indication that this is a formal parameter and therefore to be handled differently from the way you'd handle a simple variable. Some sort of scheme that would say this is a this is a, yeah, OK, OK, enough table attributes just move on

00:17:01.400 - 00:17:16.160

to something else. We only ask that question when you when you got all the uh all the at you. Well, uh, then, then there were two different ways to implement. This is the green way of implementing the displays and I I don't think I need to

00:17:16.160 - 00:17:29.199

talk about it too much. You copy it each time. Why don't you, uh, assume that you're, you're currently at this top node and you recognize it as being, uh, this an assignment node. I want you to sort of walk through it verbally describe what sort of

00:17:29.199 - 00:17:42.650

code you would, uh, generate, how you would. Um, determine the addresses of the different identifiers and, and there's 3 or 4 different issues that are implicit in this tree and sort of we'll talk about it as you go through. Uh, one part of the thing, you

00:17:42.650 - 00:18:03.800

can assume that MN X, Y, and Z are local to the procedure you're currently in. Or in your case, the block, the block, uh, the only kicker is, is that a is defined globally and Say A is at level 0 and MN X, Y, and Z

00:18:03.800 - 00:18:18.329

are at level 2. All right, OK. Well, in, in the little thought that I gave to to handling of assignments, I sort of decided it would be a good thing to evaluate the right part of the subtree first. Uh, I don't know why it just seems

00:18:18.329 - 00:18:31.719

like the order to do it in, uh, so what I would do is, is look at this. I recognize the fact that a plus operator is communicative, that you can evaluate either side of either side of the tree first, um. And as long as you can

00:18:31.719 - 00:18:42.089

do that, I might as well keep on looking at the at the uh. That that has implications too. For instance, if that was a minus node and I evaluated this much first, I'd come up with some value and then I'd want to subtract it. Well, the

00:18:42.089 - 00:18:52.709

thing to do there is to use the negate instruction, for instance, that they have rather than storing it in a temporary what all this is going to return then is an address and We remember up here we had to take this guy and we had to

00:18:52.709 - 00:19:03.390

store him in a temporary, so now he can't you talk about temporarries for a minute. Exactly how are you going to manage your temporarries? Well, it's going to be pretty inefficient since I'm not doing much optimization. I mean, all the, all the really hot algorithms for

00:19:03.390 - 00:19:16.640

temporary assignment, I can't recognize. Well, it's going to be inefficient in the sense that, you know, common subexpressions are not going to be eliminated. No, we won't worry about that. OK. What I'm really concerned with is exactly which temporary do you know to go back and

00:19:16.640 - 00:19:24.939

look in to pick up this thing that you evaluated on the right? Well, the ideal place to put it, of course, first of all, is in a register, OK? And when you start looking for stuff, you've got to see if it's. Register too because that can

00:19:24.939 - 00:19:36.060

save you time as opposed to going back to where where the copy of it is or might be in core and your registers are sort of a prime resource. You have only a very limited number of them. The safest thing on the right was was was

00:19:36.060 - 00:19:52.209

where does the temporary go? OK, well, the temporary goes right in the fixed data area, OK. For the given block. But suppose that in the process of evaluation of this, um, well, no, the M plus in here you you have a lot of more array references

00:19:52.209 - 00:20:12.180

and um more temporaries are generated and things like that. How are you going to restore things so that you get your values out of the right temporary? Well, you don't want to go to a scheme like Danzig and Reynolds, obviously, but I can't obviously not. It

00:20:12.180 - 00:20:25.819

doesn't do any good or or anything of that sort, but what would you like to? What's a neat efficient way to generate temporary so that you always know where to get the things you need, particularly for things involved in the accumulator? Well, stacks, that was, that

00:20:25.819 - 00:20:42.130

was the key buzz what we were looking for, right. Yeah OK, I, I think that, uh, covers just about all of it. You obviously have a very, very strong grasp of this, it's a good presentation, John. Yeah, like your, like your drawings. That's a good job.

00:20:43.589 - 00:20:57.300

I think engineering as a whole has got to got to turn more towards Towards building machines and systems that are gonna serve people rather than having people end up being the, the servants of the machines, for instance, here in Houston we have a, we have a

00:20:57.300 - 00:21:14.479

real problem with pollution because there's all the petrochemical industries in the ship channel. Now, in my own particular interest, the field of computer science, we've got to build machines and computers that are easier for people to work with that are easier to program. Oh. And I

00:21:14.479 - 00:21:31.609

think coming to Rice has has made me more conscious of these human aspects of engineering than if I'd gone to a school that was more, more strictly technically oriented. Well, I would, well, I mean, like what, this is only gonna be in one piece. We make

00:21:31.609 - 00:22:24.209

different kinds that match. flowers and planning. Put something underneath so that the fire wouldn't go wax paper. I don't know what you did. Now this is what my long hours in the lab. taught I had a thin out by Class where And Well, what time are

00:22:24.209 - 00:23:11.699

we gonna leave tomorrow? Well, I don't know. Whenever we feel like it. Yeah. Called the Black Canyon. Yeah. It's about 4 hours last night. We started probably 8:30. It's about 6 hours, right. No, we didn't get here till 12:30. Hm. Well, to the hatchery. So 3

00:23:11.699 - 00:23:28.239

hours. Can not canon, stop. Um, pick up that frying pan and he just want the rock so they didn't. Is that fine?