11 May 1997 · machine
11 May, New York. Six games; the decisive one lasts nineteen moves. Still the symbol of the whole subject: the machine beat the strongest chess player alive.
Deep Blue searched about 200 million positions per second and understood nothing beyond what was programmed into it.
~ August 1997 · machine
Michael Buro's Logistello beats world champion Takeshi Murakami six games to nil — a rout more complete than Deep Blue's three months earlier, and almost nobody watching.
1997 · chess · the machine won the match
The rematch, 3–11 May 1997: six games, 3½–2½ to the machine. A year earlier Kasparov had beaten the previous version 4–2.
IBM · New York, Equitable Center
The match passport: energy, price, latency, what was proved and what was not →
1997 · Othello · the machine won the match
From 4 to 7 August 1997 Michael Buro's program won all six games against the reigning Othello world champion: 6-0 in games, 264-120 in discs. It ran on an ordinary desktop machine with a single Pentium Pro under Linux.
NEC Research Institute · Princeton, New Jersey, USA
The match passport: energy, price, latency, what was proved and what was not →
IBM never disclosed what Deep Blue cost. The '$3 million' and '$5 million' figures appear in popular retellings but lead nowhere: neither Hsu's paper, nor Campbell et al., nor IBM's own page, nor Wikipedia names a sum. The only firm money around the match is the purse — $700,000 to the winner and $400,000 to the loser, with a further $100,000 to the team from Carnegie Mellon University.
Buro's report does not give a price for the machine, and none of his other texts do either. It belonged to the NEC Research Institute.
The site's table has no factor for 1995: the snapshot covers the years the matches happened in, and this price is dated to another one. The figure stays in its own money.
Intel's announced price for the 200 MHz Pentium Pro with 256 KB of L2 cache at the processor's introduction on 1 November 1995. This is the chip alone, not the machine, and by August 1997 it had fallen considerably — but no price list dated to the match could be found.
By this site's list, at the end of 1997 machines had not yet beaten the best people at 7 games. The year each of them finally fell is in brackets: Checkers: solved (2007), Jeopardy! (2011), Go (2016), Poker: heads-up (2017), Poker: six-handed (2019), StarCraft II and Dota 2 (~2019), Gran Turismo (2022).
The list holds only the games this site has a date for, and it will never be complete. Nor is the board the whole line: the abilities the site measures with its own tests — reaction, working memory, mental arithmetic and 7 more — have no crossing year at all, so they cannot be laid out by year and are not laid out here. They are gathered on The limits of a human.
Three axes on which the chips of 1997 can be compared, and who is at the top of each.
| Axis | Chip | As the catalogue has it | Compared |
|---|---|---|---|
| Most transistors | AMD K6 | 8.8M | 2 |
| Finest process | Intel Pentium II, AMD K6 | 0.35 µm | 2 |
| Highest clock | Intel Pentium II | 233–450 MHz | 2 |
There is no “best” row here. The catalogue holds no comparable measurement of performance: what its chart is drawn from are approximate ratings — reconstructed rather than measured for the older parts, spread over four different scales — and some chips in every year carry none at all. Ranking on them would display a precision the data does not have, and calling the chip with the most transistors “best” would dress a value judgement as a measurement. The three axes above are quantities with units, and their leaders disagree more often than they agree. The catalogue keeps clock speeds as a range covering a family's whole life, and the comparison above takes the top of it — a top that may belong to a later year than this page.
| Processor | Vendor | Process | Transistors | Clock |
|---|---|---|---|---|
| Intel Pentium II | Intel | 0.35 µm | 7.5M | 233–450 MHz |
| AMD K6 | AMD | 0.35 µm | 8.8M | 166–300 MHz |
Figures from this site's processor catalogue. A dash means the catalogue does not carry the value, not that it is zero.