Editing Deep Blue (chess computer) (section)

== Design ==
[[File:One of Deep Blue's processors (2586060990).jpg|thumb|One of Deep Blue's processors]]

=== Software ===
Deep Blue ran under the [[AIX operating system]], and its chess playing program was written in [[C (programming language)|C]]. 
Its [[evaluation function]] was initially written in a generalized form, with many to-be-determined parameters (e.g., how important is a safe king position compared to a space advantage in the center, etc.). Values for these parameters were determined by analyzing thousands of master games. The evaluation function was then split into 8,000 parts, many of them designed for special positions. The opening book encapsulated more than 4,000 positions and 700,000 [[Grandmaster (chess)|grandmaster]] games, while the endgame database contained many six-piece endgames and all five and fewer piece endgames. An additional database named the "extended book" summarizes entire games played by Grandmasters. The system combines its searching ability of 200 million chess positions per second with summary information in the extended book to select opening moves.<ref>{{Harvnb|Campbell|1999|page=66}}</ref>

Before the second match, the program's rules were fine-tuned by grandmaster [[Joel Benjamin]]. The opening library was provided by grandmasters [[Miguel Illescas]], [[John Fedorowicz]], and [[Nick de Firmian]].<ref>{{Cite news |last=Weber |first=Bruce |date=18 May 1997 |title=What Deep Blue Learned in Chess School |language=en-US |work=The New York Times |url=https://www.nytimes.com/1997/05/18/nyregion/what-deep-blue-learned-in-chess-school.html |access-date=4 July 2017 |issn=0362-4331 |archive-date=17 May 2017 |archive-url=https://web.archive.org/web/20170517052341/http://www.nytimes.com/1997/05/18/nyregion/what-deep-blue-learned-in-chess-school.html |url-status=live }}</ref> When Kasparov requested that he be allowed to study other games that Deep Blue had played so as to better understand his opponent, IBM refused, leading Kasparov to study many popular PC chess games to familiarize himself with computer gameplay.<ref>{{Cite web |last=Weber |first=Bruce |date=5 May 1997 |title=Computer Defeats Kasparov, Stunning the Chess Experts |url=https://www.nytimes.com/1997/05/05/nyregion/computer-defeats-kasparov-stunning-the-chess-experts.html |website=[[The New York Times]] |access-date=18 May 2020 |archive-date=24 April 2020 |archive-url=https://web.archive.org/web/20200424155122/https://www.nytimes.com/1997/05/05/nyregion/computer-defeats-kasparov-stunning-the-chess-experts.html |url-status=live }}</ref>

=== Hardware ===
Deep Blue used custom [[VLSI]] chips to [[parallel computing|parallelize]] the [[alpha–beta pruning|alpha–beta search]] algorithm,<ref>{{Harvnb|Hsu|Campbell|Hoane|1995}} p. 240</ref> an example of [[symbolic AI]].<ref>{{Cite web |last=Greenemeier |first=Larry |title=20 Years after Deep Blue: How AI Has Advanced Since Conquering Chess |url=https://www.scientificamerican.com/article/20-years-after-deep-blue-how-ai-has-advanced-since-conquering-chess/ |access-date=3 January 2022 |website=Scientific American |language=en |archive-date=20 December 2021 |archive-url=https://web.archive.org/web/20211220224453/https://www.scientificamerican.com/article/20-years-after-deep-blue-how-ai-has-advanced-since-conquering-chess/ |url-status=live }}</ref> The system derived its playing strength mainly from [[brute-force search|brute force]] computing power. It was an [[IBM RS/6000 SP]], a [[supercomputer]] with a [[massively parallel]] architecture based on 30 [[PowerPC 600#PowerPC 604e|PowerPC 604e]] processors and 480 custom [[600 nm process|600&nbsp;nm]] [[CMOS]] VLSI "chess chips" designed to execute the chess-playing expert system, as well as [[Field-programmable gate array|FPGAs]] intended to allow patching of the VLSIs (which ultimately went unused) all housed in two cabinets. The chess chip has four parts: the move generator, the smart-move stack, the evaluation function, and the search control. The move generator is a 8x8 [[combinational logic]] circuit, a chess board in miniature.<ref name="ieee-1999">{{cite journal |last1=Hsu |first1=Feng-hsiung |title=IBM's Deep Blue Chess Grandmaster Chips |journal=IEEE Micro |date=March–April 1999 |volume=19 |issue=2 |pages=70–81 |doi=10.1109/40.755469 |url=http://www.csis.pace.edu/~ctappert/dps/pdf/ai-chess-deep.pdf |archive-url=https://web.archive.org/web/20041228122752/http://www.csis.pace.edu/~ctappert/dps/pdf/ai-chess-deep.pdf |access-date=11 January 2022|archive-date=28 December 2004 }}</ref><ref name="cnet-1997">{{cite news |last1=Festa |first1=Paul |title=IBM upgrades Deep Blue |url=https://www.cnet.com/news/ibm-upgrades-deep-blue/ |access-date=11 January 2022 |publisher=Clnet |date=2 September 1997 |archive-date=3 January 2022 |archive-url=https://web.archive.org/web/20220103234251/https://www.cnet.com/news/ibm-upgrades-deep-blue/ |url-status=live }}</ref><ref>{{Harvnb|Gonsalves|2017|page=234}}</ref><ref>{{Cite book |last=Hsu |first=Feng-hsiung |url=https://books.google.com/books?id=7aBMEAAAQBAJ |title=Behind Deep Blue: Building the Computer That Defeated the World Chess Champion |date=2022-05-03 |publisher=Princeton University Press |isbn=978-0-691-23514-1 |language=en}}</ref>

In 1997, Deep Blue was upgraded again to become the 259th most powerful [[supercomputer]] according to the [[TOP500]] list, achieving 11.38 [[GFLOPS]] on the [[LINPACK|parallel high performance LINPACK]] benchmark. Deeper Blue was capable of evaluating 200&nbsp;million positions per second, twice as many as the 1996 version.<ref>{{Cite web |date=13 February 2009 |title=TOP500 List – June 1997 (201–300) |url=http://www.top500.org/list/1997/06/300 |archive-url=https://web.archive.org/web/20090213103245/http://www.top500.org/list/1997/06/300 |archive-date=13 February 2009 |access-date=10 June 2020 |website=Top 500}}</ref>