Raymond Toy pushed to branch issue-459-more-accurate-dd-complex-div at cmucl / cmucl
Commits:
-
466dee14
by Raymond Toy at 2026-01-19T11:26:01-08:00
1 changed file:
Changes:
| ... | ... | @@ -751,12 +751,12 @@ |
| 751 | 751 | (d (imagpart y))
|
| 752 | 752 | (ab (max (abs a) (abs b)))
|
| 753 | 753 | (cd (max (abs c) (abs d)))
|
| 754 | - ;; S is always multipled by a power of 2. It's either
|
|
| 755 | - ;; 2, 0.5, or +be+, which is also a power of two.
|
|
| 756 | - ;; (Either 2^105 or 2^209). Hence, we can just use a
|
|
| 757 | - ;; double-float instead of a double-double-float
|
|
| 758 | - ;; because the double-double always has 0d0 for the lo
|
|
| 759 | - ;; part.
|
|
| 754 | + ;; S is always multipled by a power of 2. It's
|
|
| 755 | + ;; multiplied by either 2, 0.5, or +be+, which is
|
|
| 756 | + ;; also a power of two. (Either 2^105 or 2^209).
|
|
| 757 | + ;; Hence, we can just use a double-float instead
|
|
| 758 | + ;; of a double-double-float because the
|
|
| 759 | + ;; double-double always has 0d0 for the lo part.
|
|
| 760 | 760 | (s 1d0))
|
| 761 | 761 | (declare (double-float s))
|
| 762 | 762 | ;; If a or b is big, scale down a and b.
|