Raymond Toy pushed to branch issue-459-more-accurate-dd-complex-div at cmucl / cmucl

Commits:

• 7fc4e1cb
  by Raymond Toy at 2026-01-10T19:13:14-08:00

  Multiply by 0.5 instead of dividing by 2
  
  For double-floats, we have transforms to convert division by 2 to
  multiplication by 0.5.  No change in precision since this is exact.
  
  However, we don't have this for double-double-floats. (We probably
  should, but that's a different issue.)  Instead, just multiply by
  0.5d0.  This means we can use the faster `mul-dd-d` routine instead of
  `mul-dd` if we multiplied by 0.5w0.
  

1 changed file:

• src/code/numbers.lisp

Changes:

 		(declare (,type s))

 		;; If a or b is big, scale down a and b.

 		(when (>= ab ,+rbig+)

-		  (setf x (/ x 2)

+		  (setf x (* x 0.5d0)

 			s (* s 2)))

 		;; If c or d is big, scale down c and d.

 		(when (>= cd ,+rbig+)

-		  (setf y (/ y 2)

-			s (/ s 2)))

+		  (setf y (* y 0.5d0)

+			s (* s 0.5d0)))

 		;; If a or b is tiny, scale up a and b.

 		(when (<= ab (* ,+rmin+ ,+2/eps+))

 		  (setf x (* x ,+be+)