BigEd wrote:
The drawback being that it's neither fast nor memory-efficient.)
Agreed but it is the easiest for realization. So if somebody wants to have the 8-bit optimization programming quiz then the spigot gives it.
The realization of a Chudnovsky's algorithm or even usage of arctangents is very much more complicated. It is curios to make a HP 50g port of spigot. Did it have assembler level programming capability? Did somebody try spigot in C with 50g? ARM evaluation system of 1986 should make 5000 digits in 280 seconds with the made spigot, IBM AT of 1984 in ≈250...
BigEd wrote:
I just want to note the section on the hpgcc page called "Viète Accelerated" which brings in an improved way to calculate with fewer square roots, by Rick Kreminski. This is referenced in this thread about computing pi on HP calculators.)
With less iterations number but with more difficulty for each iteration.
It is a big literature about this topic. Did you read a poem
https://en.wikipedia.org/wiki/Cadaeic_Cadenza? And thanks for links.
I've found an interesting link too
https://rosettacode.org/wiki/Pi. It is curious that the original work by Stanley Rabinowitz has an error in the supplied Pascal code which makes, for example, the wrong result for 1000 digits. So Bruce may make 6502 entry.
EDIT. It is not an error. The program just requires LONGINT for all variables for 1000, 3000 or 9000 digits.