# Following pi down the decimal trail.

Following pi down the decimal trail

Japanese computer scientist Yasumasa Kanada of the University of Tokyo has set himself a task that literally can never end. Step by step, he is extending the computation of pi pi -- the ratio of a circle's circumference to its diameter -- to a larger and larger number of decimal places. Earlier this year, Kanada calculated pi to 201,326,000 decimal places, shattering his own 1987 record of 134 million digits (SN: 2/21/87, p. 118). The digits of pi now known would fill every page of every issue of SCIENCE NEWS for roughly the next 28 years.

Kanada's most recent computation required 6 hours on a new supercomputer manufactured by Hitachi. He verified the result by using a slightly different computational method. Last year's effort to reach 134 million digits took nearly 36 hours on an NEC SX-2 supercomputer. The shorter time for the new calculation reflects the use of a more advanced, faster computer and the effects of tinkering with the computer program to speed it up. The basic method, or algorithm, for computing pi was not changed.

Knowing the digits of pi to millions of decimal places has little practical value. In most scientific applications, 10 decimal places are sufficient. However, for Kanada and other computer experts, computing pi is one way to test the speed and accuracy of new computers and to compare different computers. An error in even one of the millions of digits of pi would signal a problem in the computer or in the computer program.

Kanada's motivation for pursuing pi goes well beyond practical value. "It's like ~Mt.| Everest," he told SCIENCE NEWS. "I do it because it's there." His present goal is to reach 400 million digits by next year. To achieve that level, he says, he needs a computer with a greatly expanded main memory for storing the results of intermediate steps in the computation and a faster means of sending data to and from the computer.

Because pi is known to be an infinite decimal, there is no reason why Kanada cannot continue his quest indefinitely -- subject to the limits imposed by available computer technology. "I would like to go on and on," he says.

Japanese computer scientist Yasumasa Kanada of the University of Tokyo has set himself a task that literally can never end. Step by step, he is extending the computation of pi pi -- the ratio of a circle's circumference to its diameter -- to a larger and larger number of decimal places. Earlier this year, Kanada calculated pi to 201,326,000 decimal places, shattering his own 1987 record of 134 million digits (SN: 2/21/87, p. 118). The digits of pi now known would fill every page of every issue of SCIENCE NEWS for roughly the next 28 years.

Kanada's most recent computation required 6 hours on a new supercomputer manufactured by Hitachi. He verified the result by using a slightly different computational method. Last year's effort to reach 134 million digits took nearly 36 hours on an NEC SX-2 supercomputer. The shorter time for the new calculation reflects the use of a more advanced, faster computer and the effects of tinkering with the computer program to speed it up. The basic method, or algorithm, for computing pi was not changed.

Knowing the digits of pi to millions of decimal places has little practical value. In most scientific applications, 10 decimal places are sufficient. However, for Kanada and other computer experts, computing pi is one way to test the speed and accuracy of new computers and to compare different computers. An error in even one of the millions of digits of pi would signal a problem in the computer or in the computer program.

Kanada's motivation for pursuing pi goes well beyond practical value. "It's like ~Mt.| Everest," he told SCIENCE NEWS. "I do it because it's there." His present goal is to reach 400 million digits by next year. To achieve that level, he says, he needs a computer with a greatly expanded main memory for storing the results of intermediate steps in the computation and a faster means of sending data to and from the computer.

Because pi is known to be an infinite decimal, there is no reason why Kanada cannot continue his quest indefinitely -- subject to the limits imposed by available computer technology. "I would like to go on and on," he says.

Printer friendly Cite/link Email Feedback | |

Author: | Peterson, Ivars |
---|---|

Publication: | Science News |

Date: | Apr 2, 1988 |

Words: | 372 |

Previous Article: | Math society says no to SDI funding. |

Next Article: | Skeletal aging of New World settlers. |

Topics: |