Japanese and US whizzes claim news record for pi calculation -- five trillion decimal places

Aug 05, 2010

A pair of Japanese and US computer whizzes claim to have calculated pi to five trillion decimal places -- a number which if verified eclipses the previous record set by a French software engineer.

"We believe our achievement sets a new record," Japanese system engineer Shigeru Kondo said, adding that the French man's calculation to nearly 2.7 trillion places was believed to be the previous record.

The 54-year-old from central Japan, teamed up with Alexander Yee, a US computer science student, to set about calculating the constant that has fascinated mathematicians for millennia.

, the ratio of a circle's circumference to its diameter, starts with 3.14159 in a string whose digits are believed to never repeat or end.

"Alexander provided software and I was in charge of hardware. We couldn't have achieved the results without either of us," Kondo said, adding that the two men worked together while communicating by email.

It took 90 days to calculate pi at Kondo's home using a desktop computer with 20 external hard disks. It ran on the operating system Windows Server 2008R2 and used powerful Intel microprocessors. Verification took 64 hours.

Kondo built the computer by himself, procuring parts from local electronics shops and via the Internet. "I don't really want to say how much it cost me as my family may hear it... it's about 18,000 dollars," he told AFP by telephone.

It was midnight in Japan when the computer reached five trillion decimal places. "I was alone in the room at the moment... I know this is nothing but self satisfaction," he said.

His mother and wife who live with him were sleeping at that time and later showed "no particular feelings" despite his sense of achievement, he said.

Earlier this year Fabrice Bellard of France said he had used an inexpensive desktop computer -- and not a like those used in past records -- to calculate pi to nearly 2.7 trillion decimal places.

That was around 123 billion digits more than the previous set last August by Japanese professor Daisuke Takahashi.

Explore further: Professor quantifies how 'one thing leads to another'

add to favorites email to friend print save as pdf

Related Stories

Pi calculation smashes records

Jan 07, 2010

(PhysOrg.com) -- A computer scientist in France has broken all previous records for calculating Pi, using only a personal computer. The previous record was approximately 2.6 trillion digits, but the new record, ...

Hynix reports sixth straight quarterly loss

Apr 24, 2009

South Korea's Hynix Semiconductor, the world's second-largest computer memory chipmaker, Friday reported a sixth straight quarterly loss due to weak chip prices sparked by a global recession.

New computer does Windows 3,000 times faster

Jun 18, 2008

The most powerful Windows-based computer in Europe is being installed in Sweden's Umeå University. Nicknamed "Akka", the supercomputer incorporates IBM Power microprocessors, Cell Broadband Engines and Intel processors and ...

Recommended for you

Fossil arthropod went on the hunt for its prey

7 hours ago

A new species of carnivorous crustacean has been identified, which roamed the seas 435 million years ago, grasping its prey with spiny limbs before devouring it. The fossil is described and details of its lifestyle are published ...

User comments : 24

Adjust slider to filter visible comments by rank

Display comments: newest first

snowman95
2.3 / 5 (10) Aug 05, 2010
... and this proves - what? Did the embedded alien message pop put? I'm always pro-scientific inquiry, but even proving that pi doesn't repeat seems hardly worth the money spent. Maybe it is like mountaineering - just because it's there.
MustaI
1.6 / 5 (18) Aug 05, 2010
Pi is best known transcendental number so far, so you can analyze hidden structures (aether foam?) in it.

http://zenwerx.com/piorig/
http://zenwerx.com/piorig/pipic.php
sr_villarreal
Aug 05, 2010
This comment has been removed by a moderator.
MustaI
1.5 / 5 (11) Aug 05, 2010
For example, if theory of Dirac's number is valid, then the PI number should become repetitive in the range of some 40 orders.

http://en.wikiped...pothesis

Another indicia is the existence of finite order group (Moonshine group) of the size 10E+53.

http://en.wikiped...er_group
MustaI
1.6 / 5 (14) Aug 05, 2010
IMO the reliability of mainstream math and the pi number is limited by the scope of orthogonal Euclidean geometry. For larger number then the 10E+40 - 10E+55 (which roughly correspond the number of atoms inside of our universe, i.e. the number of their congruent states in human brain containing roughly 10E+23 atoms (i.e. 10E{23+23} state permutations)) the geometry of Universe cannot remain Euclidean anymore. It would explain, why we're observing the most distant parts of Universe fuzzy at both cosmological, both quantum scale: human brain simply doesn't contain enough particles for to interact with the rest of Universe in causal way.
MustaI
1.7 / 5 (12) Aug 05, 2010
Some tests of Pi number randomness...

http://home.versa...dom.html
http://news.uns.p....pi.html

For PC users it's probably faster to compute PI digits on their own, then just to download it...

http://numbers.co...ast.html
Kedas
Aug 05, 2010
This comment has been removed by a moderator.
feynmansum
5 / 5 (8) Aug 05, 2010
Mustal, have you ever considered that you may be a complete crackpot?! Dirac's number theory has nothing to do with some BS notion that it repeats at 10^53 decimals.
RobertKLR
1 / 5 (3) Aug 05, 2010
There is a fundamental problem with how we calculate the circumference and diameter ratio. For each remainder the unit of measure must be divided by ten (assuming base 10) or 100 or 1000, etc, then the process repeated, which leaves a remainder and the unit of measure must be divided again and so on. There will always be a remainder since the odds of the remainder being zero are beyond astronomical. In other words those who calculate must get lucky to find the exact value of pi. Our math is not up to the task.
bottomlesssoul
1 / 5 (2) Aug 05, 2010
Well Kondo, no matter the response of your parents it was in love. I on the other hand only heard about you from this news clip and stand in AWE of your achievement. You have seen farther than others using the rare beautiful human trait of sitting still and thinking and blessed us all by sharing it.
Jigga
1 / 5 (11) Aug 05, 2010
Dirac's number theory has nothing to do with some BS notion that it repeats at 10^53 decimals.
In dense aether yes, as this "BS notion" (i.e. sporadic gauge group) represents the closest arrangement of kissing hypersheres aka aether particles.

http://www.nature...610.html
Caliban
1 / 5 (1) Aug 06, 2010
Wonder if there is any way to correlate this with the generation of primes through iteration of the "Golden Section". I'm not sure what that would mean, but it seems like it would have some profound implications, nonetheless.
abhishekbt
1 / 5 (1) Aug 06, 2010
Thanks Mustal for the link!

Well the program suggested above seems pretty fast. I just calculated a 100M digits in just over half an hour. I am tempted to leave my PC on for a few days to see how far it will go... ;-)
Negative
3 / 5 (2) Aug 06, 2010
"Pi... starts with 3.14159 in a string whose digits are believed to never repeat or end."

believed?!?
MustaI
1.4 / 5 (11) Aug 06, 2010
"Pi... starts with 3.14159 in a string whose digits are believed to never repeat or end."
The will never repeat, but their values should exhibit a fuzzy correlations at roughly forty digits distance. You can imagine it like less or more irregular waves in landscape shape - you will never detect periodicity, when starring at last digits of your GPS, because the shape of landscape is still sufficiently random to cover such correlation.
feynmansum
4.3 / 5 (6) Aug 06, 2010
In dense aether yes, as this "BS notion" (i.e. sporadic gauge group) represents the closest arrangement of kissing hypersheres aka aether particles.


aether? kissing hyper-"sheres"? i love it.

your reference to funkhouser, even his papers referenced within the nature news article, have nothing to do with a "BS notion that pi repeats at 10^53 decimals"! how many times must I type it.

"sporadic gauge group"... perhaps you are referring to using gauge group approximations to pi? Approximations of pi are useful, have nothing to do with a miraculous prediction of a repeat at 10^53 decimals, but are related to the article above. So your jargon generator is getting warmer.

Jigga
1 / 5 (7) Aug 07, 2010
perhaps you are referring to using gauge group approximations to pi? Approximations of pi are useful, have nothing to do with a miraculous prediction of a repeat at 10^53 decimals
At 53 decimals, only. And I'm not talking about repetitions, just correlations.
KBA
not rated yet Aug 07, 2010
It's not about proving PI doesn't end (how could you compute and infinite number of places anyway!)

It is not BELIEVED that the digits of PI go on forever it has been PROVEN mathematicaly to be so. PI is irrational. There is no need to prove anything.

Parsec
4 / 5 (4) Aug 07, 2010
"Pi... starts with 3.14159 in a string whose digits are believed to never repeat or end."
The will never repeat, but their values should exhibit a fuzzy correlations at roughly forty digits distance. You can imagine it like less or more irregular waves in landscape shape - you will never detect periodicity, when starring at last digits of your GPS, because the shape of landscape is still sufficiently random to cover such correlation.

You need a course in linear analysis. Pi has been put thru every kind of statistical test known and has never been shown to have any sort of periodicity.
Parsec
3.7 / 5 (3) Aug 07, 2010
Mustal, have you ever considered that you may be a complete crackpot?! Dirac's number theory has nothing to do with some BS notion that it repeats at 10^53 decimals.

Oh please! Calling Mustal a crackpot is an insult to crackpots.

Having said that however, the desire for more digits of pi is at least in part driven by the search for some hidden periodicity. Either that or a better pseudo-random number generator.
Scalziand
not rated yet Aug 07, 2010
I thought that the point to calculating pi to ever increasing levels of precision is creating better(faster, more efficient) algorithms for it and advancing number theory. The actual calculation is just a byproduct of sorts.
MorituriMax
not rated yet Aug 08, 2010
Okay, seriously, is there an existing concrete use for knowing PI to x digits? I always thought that in mathematics, the number of digits after the decimal is the level of accuracy we are satisifed with to solve a problem.

But 5+ trillion seems like yet another useless record in todays Guiness World Record/American Idol/Attention Deficit disorder crazed culture.

Heh, but please, is there any reason to keep going higher? I am curious.
Jimster
not rated yet Aug 09, 2010
"Heh, but please, is there any reason to keep going higher?"
You will have to ask Sly and The Family Stone
Caliban
2.5 / 5 (2) Aug 09, 2010
No, ultimately, one wishes to see if Order arises out of Chaos. Is there some underlying structure in the Universe? A Music of the Spheres?

As for how high is too high- ask whoever is up there, "Is that high enough?". They'll probably reply, "I want to go higher!"
VOR
5 / 5 (1) Aug 11, 2010
maybe it reveals more about math than about the value itself. That number of decimals seems far beyond any practical application or physical reality. For example if you calculated something like the earth's orbit of the sun (pretending it was round) over its diameter, wouldn't doing so to the millimeter already be a smaller unit than would have any affect on those bodies? Im trying to say there's a limit to nature's precision, it would be interesting to know if that's true, and how that works.
SailRiver
Aug 20, 2010
This comment has been removed by a moderator.
abhishekbt
not rated yet Aug 20, 2010
OK, it's official. I calculated a billion digits of PI on my desktop PC in just over 3 days. Where do I report for my award? :-)