The Impossible Dream: Computing e to 116,000 Decimal Places...

[ledtester]

Just ran across this gem written by Wozniak in 1981:

https://archive.org/details/byte-magazine-1981-06/page/n393/mode/2up

It's very well-written and an easy to follow article. I found these closing thoughts at the end especially bemusing...

On the Horizon

As with any limitless search, there remains the challenge to compute e to even greater precision. Unfor-tunately, the computation time of the specified algorithm is exponentially related to the precision sought. Divide operations on high-speed computers (approximately 12 vs per 32 bits) are two orders of magnitude faster than the 6502 routines. The ultimate approach is to construct a custom "divide machine." Current technologies and low programmable memory prices make it feasible to con-struct such a machine with a thousand-fold performance improvement over the 6502 microprocessor. With such a machine, e could be computed to 100,000,000 digits within a couple of years (one year constructing and test-ing, one year computing). Such a machine would require power supply backup and error-correcting memory. The memory should be purchased at the latest possible date due to decreasing prices.

The current record, btw, is 31 trillion digits (2020):

http://www.numberworld.org/digits/E/

[Haenk]

I just watched pi to the pi to the pi to the pi:



That's the next level

[Ed.Kloonk]

I just watched pi to the pi to the pi to the pi:


That's the next level

42.

 

[T3sl4co1l]

Lemme just sweep all this under here...



Tim

[coromonadalix]

62.8 trillion digits - August 2021 (UAS Grisons)     

[SiliconWizard]

Lemme just sweep all this under here...
(...)

Of course the solution to determining that many decimals of irrational numbers is not to determine them in the first place, given that it has little practical value beyond a few tens of them if that.

[free_electron]

and what's the error margin ?

[Haenk]

Of course the solution to determining that many decimals of irrational numbers is not to determine them in the first place, given that it has little practical value beyond a few tens of them if that.

I doubt "a few tens of decimals" has any practical use at all
In theory, about 40 digits seem to be the maximum number to properly relate one atom to the universe:

https://www.jpl.nasa.gov/edu/news/2016/3/16/how-many-decimals-of-pi-do-we-really-need/