Precision of datatype complex

Hi,

can someone give me a answer of the precision of a complex value of matlab. is the wohle value double or each part (real,im) ?

kind regards Alex

Answer 1

[Jan]

The real and the imaginary parts are stored in a double. Therefore a complex double occupies 2*8 bytes.

[Alexander]

thanks so far! Do you know a way to store 2*16 in a complex number.

[Walter Roberson]

Symbolic Math toolbox. Fixed Point toolbox. One of the File Exchange extended precision routines. Or construct your own dataclass.

MATLAB uses hardware floating point for numeric operations. CPUs do not support 16 byte floating point numbers in hardware as yet.

Much of the time if someone needs more than 53 bit precision, they are not making realistic calculations with proper precision and round-off analysis.