Japanese Number Crunchers Calculate Pi to 10 Trillion Digits
Japanese Number Crunchers Calculate Pi to 10 Trillion Digits
Posted 10/19/2011 at 10:05am
| by Paul Lilly
10
Comments
Print
A pair of Pi lovers named Alexander J. Yee and Shigeru Kondo have calculated the mathematical constant out to 10 trillion digits. It took over a year -- 371 days -- on Kondo's desktop, not a supercomputer, to accomplish the feat, and it wasn't easy getting there. Kondo battled multiple hard drive failures, and each time an HDD would go belly up, he would have to roll back the computation to a previous checkpoint. Kondo says this added up to 180 days of lost time.
His desktop is comprised of two Intel Xeon X5680 processors running at 3.33GHz, 96GB of DDR3-1066 memory, Asus Z8PE-D12 motherboard, three LSI MegaRaid SAS 9260-8i RAID controllers, and Windows Server 2008 R2 Enterprise x64 installed on a 1TB SATA II boot drive.
Calculating Pi to 10 trillion also required about 44TB o disk space to perform the computation and another 7.6TB to store the compressed output of decimal and hexadecimal digits.
As for the last digit? It was 5, in case you're wondering. More geeky details here.
More like this
10
Comments
Comments are closed on this article
nmanguy
October 19, 2011 at 2:49pm
"As for the last digit? It was 5, in case you're wondering."
We've had the BBP formula for almost 2 decades, so we wouldn't have to calculate the nearly-ten-billion digits just to get the ten billionth digit.
thetechchild
October 19, 2011 at 9:25pm
Even the page of the project itself says so, actually... I assume the BBP is slower than the actual calculation algorithms, though, otherwise they would've used BBP. Random selection of digit blocks to verify is more likely.
On a minor offtopic note, the college student who wrote the website plays video games, does hardcore math, and has watched Clannad. +9001 geek points and a high five!
praetor_alpha
October 19, 2011 at 2:40pm
Meh, not a good benchmark, imo. And with only 40 digits or so, you can calculate the entire observable universe to within the width of a hydrogen atom; the practicality of 10 trillion digits is questionable.
biggiebob12345
October 19, 2011 at 12:27pm
Kind of a waste of time. A supercomputer would get the same result in a few hours or so by my guesstimate.
aarcane
October 19, 2011 at 12:09pm
If he would have used Linux and ZFS he would have benefitted from both GMP and the ability to use ZFS to mitigate the impact of drive failures entirely.
Wonko33
October 19, 2011 at 11:40am
What a non-story. If he did that with pen and paper then wow, but am I really impressed he left his computer running for a year? Yawn!
AETAaAS
October 19, 2011 at 11:11am
The only thing I'm interested in is not mentioned. What make were the hard drives? I would try avoiding them. :p
somethingelse
October 19, 2011 at 11:14am
Forget make of the hard drive, what kind of raid level were these guys running? Three raid controllers, and each time a drive popped, there'd be data loss? Was he running raid0 arrays? With that kind of gear, raid5/6 would not add a lot of overhead. raid0+1 would be better, that would eat up a lot more space.
Unless it's the 2TB OS drive that kept popping, but then why would they need to roll back?
As for how this benefits anyone...well, it doesn't...who cares :)
d3v
October 20, 2011 at 2:23pm
It says they wrote 44tb of data during the calculations which was later compressed down to 7.6tb. I imagine all those writes took their toll on the hard disks and that is why they died.
iceman08
October 19, 2011 at 10:19am
And this benefits me...?
Props for using a desktop, but I'd rather eat pie than know pi
Connect with MaximumPC
Featured Content
The gang discusses the GTX 690, Trinity, Ivy Bridge, Amazon, big-box stores, MMOs,...
This month's issue
Feature
Build a PC On Any Budget
Feature
Cloud Services Showdown
How-To
Encrypt Your External Drive
Build It
Upgrade an X58 Rig
Most Commented Articles
Latest Max PC Tweets
- maximumpc: Original RickRoll Video Returns to YouTube After 24-Hour Copyright Hiatus http://t.co/MaH9Sxyy1 day 16 hours ago
- maximumpc: Corsair has decided not to go forward with its $78 million IPO, citing weak equity market conditions http://t.co/A5a4DAzj1 day 16 hours ago
- maximumpc: Want to work at Maximum PC? Our Online Managing Editor, Alex, is moving away, so we need a new one. http://t.co/9Z7JCh0E3 days 13 hours ago
