Re: [julia-users] Re: Informal Call for a Julia π - day challenge

[Hans W Borchers]

I can provide an implementation of the "droplet" algorithm for computing 
pi, based on work by Borwein, Rabinowitz, and Wagon. It does neither use 
'mpfr' nor big floats, only simple arithmetic. It will spit out about 
50,000 digits one after the other -- with more digits it gets too slow. A 
speed-up of the factor 10 or so will be possible with some tricks.
As a test I am abut to reimplement it in Nim. As Nim can also generate C or 
C++ code, it will be interesting to compare C, Nim, and Julia on this 
non-trivial, but straightforward algorithm.