Sollya User's Manual

Name:

honorcoeffprec indicates the (forced) honoring the precision of the coefficients in implementpoly

Library names:

sollya_obj_t sollya_lib_honorcoeffprec() int sollya_lib_is_honorcoeffprec(sollya_obj_t)

Usage:

honorcoeffprec : honorcoeffprec

Description:

Used with command implementpoly, honorcoeffprec makes implementpoly honor the precision of the given polynomial. This means if a coefficient needs a double-double or a triple-double to be exactly stored, implementpoly will allocate appropriate space and use a double-double or triple-double operation even if the automatic (heuristic) determination implemented in command implementpoly indicates that the coefficient could be stored on less precision or, respectively, the operation could be performed with less precision. See implementpoly for details.

Example 1:

   > verbosity = 1!;
   > q = implementpoly(1 - dirtysimplify(TD(1/6)) * x^2,[-1b-10;1b-10],1b-60,DD,"p","implementation.c");
   Warning: at least one of the coefficients of the given polynomial has been rounded in a way
   that the target precision can be achieved at lower cost. Nevertheless, the implemented polynomial
   is different from the given one.
   > printexpansion(q);
   0x3ff0000000000000 + x^2 * 0xbfc5555555555555
   > r = implementpoly(1 - dirtysimplify(TD(1/6)) * x^2,[-1b-10;1b-10],1b-60,DD,"p","implementation.c",honorcoeffprec);
   Warning: the inferred precision of the 2th coefficient of the polynomial is greater than
   the necessary precision computed for this step. This may make the automatic determination
   of precisions useless.
   > printexpansion(r);
   0x3ff0000000000000 + x^2 * (0xbfc5555555555555 + 0xbc65555555555555 + 0xb905555555555555)

See also: implementpoly, printexpansion, fpminimax