Name:
worstcase
searches for hard-to-round cases of a function
Library names:
void sollya_lib_worstcase(sollya_obj_t, sollya_obj_t, sollya_obj_t,
sollya_obj_t, sollya_obj_t, ...)
void sollya_lib_v_worstcase(sollya_obj_t, sollya_obj_t, sollya_obj_t,
sollya_obj_t, sollya_obj_t, va_list)
Usage:
worstcase(function, preimage precision, preimage exponent range, image precision, error bound) : (function, integer, range, integer, constant) -> void
worstcase(function, preimage precision, preimage exponent range, image precision, error bound, filename) : (function, integer, range, integer, constant, string) -> void
Parameters:
- function represents the function to be considered
- preimage precision represents the precision of the preimages
- preimage exponent range represents the exponents in the preimage format
- image precision represents the precision of the format the images are to be rounded to
- error bound represents the upper bound for the search w.r.t. the relative rounding error
- filename represents a character sequence containing a filename
Description:
- The worstcase command is deprecated. It searches for hard-to-round cases of
a function. The command searchgal has a comparable functionality.
Example 1:
> worstcase(exp(x),24,[1,2],24,1b-26);
prec = 165
x = 1.99999988079071044921875 f(x) = 7.3890552520751953125 eps = 4.5998601423446695596184695493764120138001954979037e-9 = 2^(-27.695763)
x = 2 f(x) = 7.38905620574951171875 eps = 1.44563608749673018122228379395533417878125150587072e-8 = 2^(-26.043720)
Go back to the list of commands