Marko Vojinovic wrote:
On Sat, Nov 6, 2010 at 8:43 AM, Mogens Kjaer mk@lemo.dk wrote:
http://www.kurims.kyoto-u.ac.jp/~ooura/pi_fft.html
You'll need RAM to get many digits.
1.6 G decimals in 20 hours on a machine with 16G RAM, running x86_64 Fedora 12.
Really, I'm curious, is there any real-world problem where anyone would actually *need* pi to a G decimal places? I mean, are these kind of computations actually useful for someone, or is it just a matter of "we have the power to do it, so let's do it" thing? Other than entry into the Guinness book of records, that is?
Yes, there is. Belated response, I know.
There are two important uses for such computations I can think of off the top of my head.
First, it's a good test of the functionality of a new machine. When the first new unit runs off the factory floor, this type of lengthy computation with known results is a good test.
Second, there are certain theoretical ideas about the distribution of digits in transcendental numbers which can make progress via such lengthy computations.
Nobody needs more than about 6 figures for doing any engineering work, however.
Mike