_ - - -- -e -__t_ v-.- .‘__. - v- ..u "no" Wye-v. VII“: UCIIIIUI W Will in war: microelectronics lab; it must be manufactured in Bell-Northern research labs, in Ottawa, or in other high tech labs in California. The students are (left to right): Jim Jonkman, Hamilton, electrical engineering; Tom Schultz, Wisconsin. computer science; Greg Bakker, formerly of Hamilton, electrical engineering. and Schaeffer. computer scrence. - . - - _ _ - - - 7 C-T-r----'-"---"--"-"--'----- CHESS CHIP DESIGNERS. Four University of Waterloo graduate students have been designing a microchip to vastly speed-up a computer chess program originally developed by Jonathan Schaetter (right). The students have been using sophisticated, very large scale integration (VLSl) technology in the design of the chip, which will have about 1 5,000 transistors on a device one-quarter of an inch square. Chip cannot be built in IIW‘- _iA--egl--b_-_:‘- I-|._ " _,,-. L _ - - - - _ H - - The chip has built into it virtually all the legal moves with re- spect to any piece on any square on the chess board. "Basically, what the ?LMG does is take the move-generation func- tion out of the soft, ware," says Schaeffer. “In other words, we build it into a piece of hardware - a chip. Thus the computer has to spend virtually zero time on this function. It vastly simplifies the computer's job." Schaeffer and his as- sociates; Tom Schultz, also a computer science grad student. and Jim Jonkman and Greg Bakker, electrr. cal enginnering stu- dents, have designed a new microchip. This is a tiny silicon wafer containing about 15,000 transistors: it is called a "pseudo-legal move generator" (PLMG) for the game of chess. When Jonathan Schaeffer, a PhD com- puter science student at the University of Waterloo. failed to win the North American computer chess chant- pionship with his "Prodigy" computer chess program a little over a year ago, he wasn't discouraged. He regarded the failure as a challenge. As a result, he and three other UW gradu- ate students have de, veloped a new piece of computer technology tor at least are well along in the process of developing it) which may just. some day, permit a computer to play the best chess in the world. NEW YEARS EVE SllLllEllllEllPotnlllat, Chronicle Special Friday, December 31, 1982 q - " your old: The highlight of the year! Our program includes games, _swimming, trampoline and parachute. a magic show, cartoons, a feature movie and a huge New Year's Eve Celebration! Snacks and a full hot breakfast will be sewed. The children should arrive at 6:00 pm. on Friday night and be picked up at 1 1:00 am. on Saturday morning. UW grad students attempt to create the world’s best chess player Waterloo Family Y's FAMOUS Cost: $21.00 one child $35.00 two children (same family) $42.00 three children (same family "One simply can't buy anything remotely like our chip on the market," Schaeffer says. The four UW stu- dents feel their chip represents a signifi- cant advance over the Bell chip. Even at that they have been told the United States govern- ment has classified the latter to make sure it does not pass into the hands of Some un- friendly foreign gov- ernment. He says the Bell Labs microchip PLMG uses either medium-scale integration (MSI) or large-scale integration (LSI) but the one the four UW students have designed is based on very large scale inte- gration (VLSI) which means it will be smaller and more com- pact - and therefore faster - than the Bel Lab chip. "One advantage of this was they could look ahead many moves into the future and check Gut all sub- sequent possibilities for each move," says Schaeffer. "We were unable to look ahead nearly as far." Schaeffer says that during the chess com- petition late in 1981 he learned there was nothing inherently wrong with his Prodigy program. It just ran too slowly. it would analyze about 50 chess positions per second; Bell Labs of New Jer- sey, had a computer chip that functioned as a PLMG which permit- ted the computer to analyze about 130,0o0 chess positions per sec- BIGISTIR TODAY 145 Lincoln Rd. Waterloo 885-3 soo Schultz is a native of Wisconsin. He is an employee at Bell Lab., N.J. His employer is sponsoring his educa, tion at Waterloo where he is studying for a master's degree. Schaeffer is from Thornhill. He received his bachelor's degree from the University of Toronto and has com- pleted a master's de, gree in computer science at Waterloo. Bakker and Jonkman are from the Hamilton area. They are electrr cal engineering gradu- ates of UW and are on master's programs. "Once we have the chip we will have to find out how to make it work," says Schultz. "We don't expect it to work the first time; still, our hope is that it can be made to work before very long. We are satisfied that the design is sound." The initial chip would involve 16 chess squares, (a regular chessboard has 64 squares). The instruc- tions will be in the form of numerical data which will enable a computer to control the preparation of the masks tor the com- puter chip. These will be reduced many times, photographical- ly, for etching onto a silicon chip. The actual chip will be about a quarter of an inch square. The rirst stage of the design has been com- pleted and the students have been working on a series of instructions which will enable Belr Northern Research Ltd., Ottawa, to actual. ly produce the chip.