1 / 44

CENG 471

CENG 471. PARALLEL COMPUTING FALL 2007 Instructor: Y. Doç. Dr. Cem ÖZDOĞAN. DEEP BLUE THE CHESS PLAYING SUPERCOMPUTER. PARALLEL COMPUTING.

lewisharold
Download Presentation

CENG 471

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CENG 471 PARALLEL COMPUTING FALL 2007 Instructor: Y. Doç. Dr. Cem ÖZDOĞAN

  2. DEEP BLUE THE CHESS PLAYING SUPERCOMPUTER

  3. PARALLELCOMPUTING

  4. Parallel computing is the use of two or more processors (computers) in combination to solve a single problem. The programmer has to figure out how to break the problem into pieces, and has to figure out how the pieces relate to each other.

  5. A good exampleof parallelism is automobile manufacturing. In automobile manufacturing construction of the automobile components are done in parallel. The engine, headlights, body, drive train, etc., are built by different companies simultaneously and their output is coordinated to arrive at the assembly plant at the appropriate time.

  6. The task of manufacturing the components for a single car by one person is not really within the realm of possibilities. Breaking this task into smaller tasks and assigning them to different companies makes what was an impossible task into one that is quite manageable. By the same token, a complex program can be broken down into smaller tasks and distributed among different processors to perform the same task faster and more efficiently.

  7. For example, a parallel program to play chess might look at all the possible first moves it could make. Each different first move could be explored by a different processor, to see how the game would continue from that point. At the end, these results have to be combined to figure out which is the best first move. Actually, the situation is even more complicated, because if the program is looking ahead several moves, then different starts can end up at the same board position.

  8. To be efficient, the program would have to keep track of this, so that if one processor had already evaluated that position, then others would not waste time duplicating the effort. This is how must parallel chess-playing systems work, including the famous IBM Deep Blue machine that beat Garry Kasparov.

  9. DEEP BLUE

  10. The computer system "Deep Blue" was the first machine to win a chess game against aworld champion, Garry Kasparov under regular time controls. The project was started as "Chip Test" at Carnegie Mellon University by Feng Hsuing Shu; Hsu joined IBM in 1989 and worked with Murray Campbell. They were to explore how to use parallel processing to solve complex computing problems. Deep Blue was developed out of this. The name is a play on Deep Thought and Big Blue(Common nickname of IBM).

  11. Deep Thought was a computer designed to play chess. It was second in the line of chess computers developed by Feng Hsuing Hsu, starting with Chip Test and resulting in Deep Blue. Deep Thought was easily defeated in both games of a 2-game match with Kasparov in 1989. It was named after Deep Thought, a fictional computer in Douglas Adams' series, The Hitchhikers Guide to the Galaxy. The naming of chess computers has continued in this vein with Deep Blue, Deep Fritz(German Chess Program), Deep Junior(Israeli Chess Program), etc. ("Deep" here generally refers to the special ability to use multiple processing units)

  12. DEEP BLUE - 1

  13. Deep Blue I was based on a single-chip chess search engine, designed over a period ofthree years. The first chips were received in September of 1995. A number of problemswere found with these chips, and a revised version was received in January of 1996. Deep Blue I ran on a 36-node IBM RS/6000(RS/6000 is IBM’s current RISC(Reduced Instruction Set Computer)/UNIX based server) computer, and used 216 chess chips.The chips each searched about 1.6–2 million chess positions per second. Overall searchspeed was 50–100 million chess positions per second.

  14. The full 36-node Deep Blue I played only six games under tournament conditions, allin the February 1996 match with Garry Kasparov. This match was won by Kasparov bya fairly decisive 4–2 score, although the match was tied at 2–2 after the first four games. • Three additional tournament-condition matches were played in preparation for the 1996Kasparov match, all using a single-node version of Deep Blue with 24 chess chips. This system, beat Grandmaster Ilya Gurevich 1.5–0.5, drew GrandmasterPatrickWolff 1–1, and lost to Grandmaster Joel Benjamin 0–2.

  15. DEEP BLUE - 2

  16. After the 1996 match with Kasparov, it wasclear that there were a number of deficienciesin the play of Deep Blue I.A series of changes were made in preparation for the rematchwhich took place in May of 1997.First, a new, significantly enhanced, chess chip wasdesigned. The new chess chip had a completely redesigned evaluation function, goingfrom around 6400 features to over 8000.The new chip also added hardware repetition detection, anumber of specialized move generation modes, and some efficiency improvements that increased theper chip search speed to 2–2.5 million positions per second.

  17. The second major change wasto more than double the number of chess chips in thesystem. A third changewas the development of a set of software tools to aid in debugging and match preparation. Finally, we concluded that the searchingability of Deep Blue was acceptable, and we spent the vast majority of our time betweenthe two matches designing, testing, and tuning the new evaluation function.

  18. Deep Blue defeated Garry Kasparov in the 1997 match by a score of 3.5–2.5. Forthis victory, the Deep Blue team was awarded the Fredkin prize for defeating the humanworld champion in a regulation match. There were two additional matches played by DeepBlue Jr. in preparation for the Kasparov match. The two matches, against Grandmasters1 This last match went a long way to convincing us that Joel Benjamin would be an excellent Grandmasterconsultant to the Deep Blue team.Larry Christiansen and Michael Rohde, were both won by Deep Blue Jr. by a scoreof 1.5–0.5.

  19. ABOUT DEEP BLUE

  20. Deep Blue is massively parallel(A massively parallel computer is a distributed computer system which consists of many individual nodes, each of which is essentially an independent computer in itself, and in turn consists of at least one processor, its own memory, and a link to the network that connects all the nodes together. Nodes communicate by passing messages, using standards such as MPI. ), 30 node, RS/6000. • Deep Blue has 480 special purpose VLSI(Very Large Scale Integration - is the process of creating integrated circuits by combining thousands of transistor-based circuits into a single chip. ) chess chips.

  21. Its chess playing program was written with C. • It runs under AIX operating system. AIX stands for Advanced Interactive eXecution and it was developed by IBM.Its source code is closed and operating system family is UNIX System V. • It had 32 processors.(Ten years later, Blue Gene, the fastest supercomputer in the world and the descendent of Deep Blue, uses 131,000 processors to routinely handle 280 trillion operations every second. ) • It has a system, capable of examining 200 million moves per second, or 50 billion positions, in the three minutes allocated for a single move in a chess game.

  22. It was designed to play chess at the grandmaster level. • Deep Blue does not use any artificial intelligence. There is no formula for chess intuition. Deep Blue relies on computational power and a search and evaluation function. • Deep Blue is the first computer to attain celebrity status and ended up on the cover of Newsweek ("The Brains Last Stand"). Deep Blue is recognized by over 50% of all Americans.

  23. In June 1997, Deep Blue was the 259th most powerful supercomputer, capable of calculating 11.38 gigaflops, although this did not take into account Deep Blue's special-purpose hardware for chess. • IBM spent 5 years and millions of dollars building Deep Blue. • The Deep Thought/Deep Blue team consisted of Hsu, Campbell, Joe Hoane, Jerry Brody, and C.J. Tan.

  24. DEEP BLUE TEAM

  25. Jerry Brody, an engineer, made adaptations on the machine hardware. Joel Benjamin, a grandmaster, joined the team full-time in August, in the latter stages of the chip design. He spent his time playing position after position against Deep Blue, filling up notebooks with observations and trying to locate as many situations as possible in which the computer evinced weaknesses.

  26. Hoane then had the responsibility of translating Benjamin's observations into computerspeak. Murray Campbell, the team member with the greatest chess expertise, was responsible for integrating chess knowledge into the system. And Feng designed the new accelerator chip.

  27. GARRY KASPAROV

  28. Garry Kimovich Kasparov was born in Baku, Azerbaijan,in 1963. Kasparov learned to play chess from his father who later died in a road accident when he was 7 years old. He changed his name to Kasparov, a Russified version of his mother's maiden name, Kasparyan. Kasparov's chess talent was apparent at an early age. In 1973 he attended the Botvinnik Chess School and Kasparov continued to make rapid progress. In 1975 at the age of 12 he became the youngest ever player to win the USSR Junior Championship. At 16 he won the World Junior Championship. He achieved the title of Grandmaster on his 17th birthday.

  29. In 1996 Kasparov competed in a six game match againstDeep Blue. Kasparov won with a score of 4 games to 2 games.  The following year, he competed again and was defeated 3.5 games to 2.5 games. It was the first time a Grandmaster had lost a series of games to a computer. He is currently the highest rated player there has ever been in the history of chess.

  30. KASPAROV vs. DEEP BLUE

  31. Deep Blue first won fame by playing Garry Kasparov in Feb, 1996. Deep Blue lost that time, but an improved model was ready for him in 1997. The improved version had twice the processor speed and the 1996 version. In the 1996 match, Deep Blue was able to win game 1. For the first time under tournament conditions, the world chess champion had lost to a computer. Kasparov eventually won 3 games and drew 2 games to win the match.

  32. In May, 1997 Deep Blue (affectionately called Deeper Blue) played world chess champion Garry Kasparov and won the match. Deep Blue won 2 games, lost 1 game, and drew 3 games to take the match in 6 games. The match was observed by 5 million people over the Internet. Kasparov accused IBM of cheating and demanded a rematch, but IBM declined and retired Deep Blue.

  33. FUTURE

  34. Feng Hsuing Hsu later claimed in his book Behind Deep Blue: Building the Computer that Defeated the World Chess Champion that he had the rights to use the Deep Blue design to build a bigger machine independently of IBM to take Kasparov's rematch offer, but Kasparov refused a rematch. *http://www.chesscenter.com/twic/feng.html The link of Feng Hsuing Hsu’s open letter.

  35. Kasparov's side responded that Hsu's offer was empty and more of a demand than an offer because Hsu had no sponsors, no money, no hardware, no technical team, just some patents and demands that Kasparov commit to putting his formal world title on the line before further negotiations could even begin.

  36. Kasparov's loss to Deep Blue inspired the creation of a new game called Arimaa, which is still played with a standard chess set, but which is also thought to be much more difficult for computers. Deep Blue, with its capability of evaluating 200 million positions per second, was the strongest computer that ever faced a world chess champion. Today, in computer chess research and matches of world class players against computers, the focus of play has often shifted to software chess programs, rather than using dedicated chess hardware.

  37. Modern chess programs like Rybka, Deep Fritz or Deep Junior are more efficient than the programs during Deep Blue's era. In a recent match, Deep Fritz vs. Vladimir Kravnik in November 2006, the program ran on a personal computer containing two Inter Core 2 Duo CPUs, capable of evaluating only 8 million positions per second.

  38. REFERENCES http://www.geocities.com/SiliconValley/Lab/7378/deep.txt http://www.research.ibm.com/deepblue/meet/html/d.4.shtml http://www.rci.rutgers.edu/~cfs/472_html/Intro/NYT_Intro/ChessMatch/WhatDeepBlueLearned.html http://en.wikipedia.org/wiki/IBM_Deep_Blue

  39. REFERENCES http://www-03.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV1001.html http://www.math-info.univ-paris5.fr/~bouzy/Doc/PJR/DeepBlue.pdf http://www.chesscorner.com/worldchamps/kasparov/kasparov.htm

  40. ANY QUESTIONS ?

  41. THANK YOU FOR LISTENING ME !

More Related