Mathematicians Calculate 10 Trillion Digits of Pi With Xeons

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

mb2bm55

Distinguished
Nov 16, 2010
55
0
18,630
[citation][nom]alidan[/nom]
actually, this is a complete waste.remember, pi is used to calculate how round a circle is. in reality, there is no such thing as a perfect circle, and that number would end, but this is a theoretical number. take a look at a cars pistons, as a common example, i beleive that we are more than close enough there, and in a more scientific view, they always account for a margin of error. lets say this was needed land something on mars, the margin of error that they calculate would make such a precise number pointless. this is a pointless waste of resources, and almost any application of the power would have been better.[/citation]

Actually it is of enormous importance in the philosophy of physics and mathematics. If pi and alternatively e are in fact proven to be irrational non-repeating and transcendental (to the most significant degree we can, remember you can only prove a number is not irrational non-repeating. Hence the title) it would have seismic impact that would ripple out through the mathematics and physics fields. These properties underlie some of the most important arguments in the fields such as the existence of purely probabilistic effects (as opposed to merely large scale statistical). This of course is important to our very approach to Quantum Mechanics and therefore underlies all of the scientific and engineering fields. Besides math is the study of proportionality, and physics is the study of the perceived universal proportionality. pi is one of the 5 or 6 cornerstone constants along with 0, 1, i and e (along with c , pending the opinion of cern's neutrino's). So, obviously, the properties of pi are rather important.
 

kristoffe

Distinguished
Jul 15, 2010
153
12
18,695
I am pretty sure they're not morons. The point of pi at such lengths it maybe to use it for patterning with encryption. set the point, use a certain set of digits, encode, decode. I think it is interesting to read about, links to pi itself and the threads on it's number theory would be nice:

http://en.wikipedia.org/wiki/Pi
 

phyco126

Distinguished
Nov 6, 2011
1,014
0
19,460
I love how people are calling this a complete waste of time and power. Right. So what exactly does Facebook, games, and porn accomplish? Absolutely nothing worth the time and power wasted.
 

blader15sk8

Distinguished
Feb 6, 2006
232
1
18,715
[citation][nom]intel4eva[/nom]retardeda 2600K at 5Ghz does 1 million digits of pi (using super pi) in less than 8 seconds.8 seconds ~ 1Million digits80000 seconds ~10Billion digits80000 seconds is approximately 22 hours.This is on one modern processor overclocked to 5Ghz which is doable on air.The question is what took these morons 371 days? (longer than a year)[/citation]

The time to calculate the digits of Pie does not scale linearly, but rather exponentially.
 

billcat479

Distinguished
Mar 19, 2006
74
0
18,630
Ah, Abstract knowledge, the drive to know everything.
There is no limit to the amount of information we have stashed away and while some MAY never be used some of it may just solve problems that we haven't had to face yet. Or it may be critical to extending our existence into the future by another hundred years to a billion.
A good lesson on the disaster's of lost information through out history is a hard lesson indeed. The burning of the huge centers of information that happened from long ago wars has erased so much information that helped keep us in the dark ages a lot longer than it could have been after Rome was sacked.
We could have been light years further than we are now because of lost information.
Most of mankind has a drive to know everything and while it seems stupid to most it is a lot more stupid to quit thinking don't you think? Or not think? Or can't think?
Or as most posts show most aren't thinking past the short term limits of their imagination.
What if this number was useful to building the first working Fusion power plant? Or working out a way to divert a asteroid from hitting the earth in 300 years with a attached rocket motor and needed the accuracy to know for sure it would work and know it's path for the next 10,000 years. The reasons for abstract knowledge is seldom known at the time but is there for the time to be used in one form or another when the need arises.
Like the probes they send to other planets, most of what they learn may seem pointless like some thought about sending them to Venus but we learned a lot from it even if we never learn to apply it wisely. Global warming is one thing we learned from it.
We know what the earth may be like if we don't find a way to stop pumping megatons of chemicals into our planets air. At least we will know how we are killing ourselves and will have a good deal of data that will tell us when the point of no return is going to be passed when it becomes a runaway process that heats up the planet enough to boil the water into the air making the planet uninhabitable. And even know how many hundreds of thousands of years it will take to cool back down to where it's able to support life again.
Fun huh?
We thrive on the abstract data acquisition. It's what makes us what we are. It's what is used to go past levels of technology we may think is impossible today.
It's not stupid to do these things. The stupid part is to quit doing it and regress back to the dark ages because of any lack of trying to be better informed to avoid it.
If we ever get space based industry going they may be able to use numbers that high, or higher in industry. That would be huge.
Even doing it for simple ball bearings would be a milestone. Build machines that are almost free from wear, engines that may last for thousands of years? A bit of maybes but you never know how things play out as computers get better and we learn more about the atomic structure and work in that arena.
It's almost painful to think any data being thought of as stupid or useless but you have to be able to think outside of limitations to see the value of it.
 

cybersans

Distinguished
Dec 28, 2009
74
0
18,630
nowadays, so-called dekstop processors can beat so-called server processors is many things. it is a waste to build a expensive hardware to run small scale server farm. just buy latest 2nd generation i7 and build your server. it is cost effective, less power draw, and kick more ass comparing to xeon in several things.

p/s: to amd fan boys, go play somewhere else. don't you read the topic before making a posting? is there state the word AMD? :p
 

kristoffe

Distinguished
Jul 15, 2010
153
12
18,695
@cybersans, I agree with you that the i7^2 is amazing and getting the benefits of what we previously considered server only. I have always scoffed at my employers for purchasing high end equipment, starting off with 32meg opengl cards in 1997 for $2400, when the ATI radeon series back then had 4-8mb and did almost exactly the same thing with hacked drivers, even out of the box.

but for specific jobs like incredible multithreading and specific calculations, intel hides some special sauce in those xeons and overcharges like a mofo. Some calculations I believed to be GPU consumable are not geared towards parallel processing. I found a special approach to pi after I rad this article in wiki that resolves 14 digits per calculation, ergo a form of parallel processing in a metaphorical sense.

I want an i7^1~^2 that can have a dual processor on a board, but for almost anything beyond 3d raytracing I cannot find a purpose, even H264 encoding video streams for previz and forensics can be done with the 3000 series built into the i7 or a series of NVIDIA cards (yes opencl can do it if programmed well as Open Source projects show). But for mathematica and other projects, it still seems that cpus are geared towards calculations beyond gpus in certain approaches...

as are these high layered and leveled buffer xeons. If intel would just spread the special sauce on the i7 series and allow everyday joe to do this as well maybe even with a hack here and there we could all benefit. I think they don't do this for the throw away and cyclical buying factor. keep people employed by geoverning otherwise 100% scientific processors and cattling them into 50+ versions. (lame).
 

back_by_demand

Splendid
BANNED
Jul 16, 2009
4,821
0
22,780
tumblr_lugilnbhBG1qhkgm0o1_500.jpg
 
G

Guest

Guest
Great work! It's neat to see someone doing this from a home (though server-class) computer out of their basement.

While I'm sure it will have mathematical implications, I doubt it will make anything more precise in areas of mechanics and tolerance levels. The circumstance of the known universe (the largest known "thing", save multi-verse theories) can be calculated to within the size of a hydrogen atom with 39 digits of pi. According to PBS/Nova, the "strings" of string theory (what we can conceptualize as the smallest known "thing) are about 10^-33 centimeters long.

That means you can calculate the circumference largest known thing in units of the smallest known thing with about 33 + 39 = 72 digits of pi (actually subtract a few, as a centimeter is quite a few scales larger than an atom - but what's a few orders of magnitude).

72 digits. We now know 10,000,000,000,000 (10 trillion) digits of pi.

Or, put a different way, almost 2/3 of the US national debt (in US dollars).
 
G

Guest

Guest
what, wait? a 1 with 13 zeros um 1 0000000000000 challenge accepted.
 

mas6700

Distinguished
Nov 10, 2011
5
0
18,520
Actually Pi to something around 40 digits will allow you to calculate the circumference of a circle the size of the orbit of Jupiter to within an accuracy of the diameter of a Hydrogen atom. So it seems to me that all those Xeons could be put to better use other than calculating an infinite number that has no practical use.
 

wardler

Distinguished
Jun 28, 2011
91
0
18,630
[citation][nom]alidan[/nom]actually, this is a complete waste.remember, pi is used to calculate how round a circle is. in reality, there is no such thing as a perfect circle, and that number would end, but this is a theoretical number. take a look at a cars pistons, as a common example, i beleive that we are more than close enough there, and in a more scientific view, they always account for a margin of error. lets say this was needed land something on mars, the margin of error that they calculate would make such a precise number pointless. this is a pointless waste of resources, and almost any application of the power would have been better.[/citation]
I was going to downrate the guy too... but then I read the last line and crapped my pants it was so funny.
 

fomoz

Distinguished
Oct 22, 2010
87
0
18,640
[citation][nom]back_by_demand[/nom]Are they hoping that when they get to something daft like 100 trillion digits is stops recurring and becomes a fixed number?[/citation]
No, because they know that Pi is a transcendental number.
 
Status
Not open for further replies.