Hello there,
I am looking for an algorithm i once had long ago. I havent used it in maybe 10 years or longer although i did post it here somewhere at least once but it could have been in a PM. I hope it wasnt though.
This algorithm calculates the sine of an angle, ti's that simple, but it does so very efficiently because it doubles the digits of precision with every pass. So if the exact answer was 0.123456780 and we started with 0.12, the next pass would produce 0.1234... and the next pass would produce 0.12345678 (doubling the correct digits with each pass).
By comparison, Taylor's and other series converge rather slowly, too slowly really, so i wanted to find the better algorithm again. The function would simply calculate the sine of an angle sin(x) and it was a fairly simple function with maybe add, subtract, multiply, divide, maybe squaring and maybe square root, but it was quite short too not long and complex.
Note it is also not an identity type formula like 3*sin(x/3)-4*sin(x/3)^3 which also works but is slow to converge.
Any help finding this either here or somewhere else would be a lot of help.
Thanks much.
I am looking for an algorithm i once had long ago. I havent used it in maybe 10 years or longer although i did post it here somewhere at least once but it could have been in a PM. I hope it wasnt though.
This algorithm calculates the sine of an angle, ti's that simple, but it does so very efficiently because it doubles the digits of precision with every pass. So if the exact answer was 0.123456780 and we started with 0.12, the next pass would produce 0.1234... and the next pass would produce 0.12345678 (doubling the correct digits with each pass).
By comparison, Taylor's and other series converge rather slowly, too slowly really, so i wanted to find the better algorithm again. The function would simply calculate the sine of an angle sin(x) and it was a fairly simple function with maybe add, subtract, multiply, divide, maybe squaring and maybe square root, but it was quite short too not long and complex.
Note it is also not an identity type formula like 3*sin(x/3)-4*sin(x/3)^3 which also works but is slow to converge.
Any help finding this either here or somewhere else would be a lot of help.
Thanks much.