On Sat, May 28, 2016 at 3:48 PM, digxx <diger_d...@hotmail.com> wrote: > Well thx yes this works. But it is not the notation I found at: > http://fredrikj.net/arb/acb_hypgeom.html#convergent-series > > Also I presume R=AcbField(128) set the float to 128. Does it always consume > the 128 bit for real and 128 for complex coz it always adds the complex part > for example R(1) =1.000.... + i*0
Yes, this is to use complex arithmetic with 128 bits of precision for both the real and imaginary part. As long as the imaginary part is exactly 0, it is functionally equivalent to 128-bit real arithmetic. > Does it change the imaginary part from im to i ? Im wondering coz for > example R(1+1*im) does not work. The interface doesn't support conversions from all Julia types; you can use R(1,1) to create a complex number from given real and imaginary parts. > I was also wondering why u need to write the values witihin hyp1f1 in > AcbField format not as normal Float128 for example which Julia provides. > Is there a documentation? Thx. This is because the function is intended for use with Arb/Acb numbers. You can use it with Julia floating-point numbers by doing the conversions float -> Acb -> hyp1f1 -> Acb -> float, perhaps writing a convenience wrapper function for this. It would be certainly be very nice to have an interface that does this automatically, but no one has had time to write it yet. There's documentation here: http://nemocas.org/nemo-0.4.pdf Best, Fredrik
