diff options
| author | Szabolcs Nagy <nsz@port70.net> | 2013-05-19 14:43:32 +0000 |
|---|---|---|
| committer | Szabolcs Nagy <nsz@port70.net> | 2013-05-19 14:43:32 +0000 |
| commit | 1e5eb73545ca6cfe8b918798835aaf6e07af5beb (patch) | |
| tree | d1af43fb6a86387586e5fbdcede7d2a3f56c78d4 /src/math/fmal.c | |
| parent | ffd8ac2dd50f99c3c83d7d9d845df9874ec3e7d5 (diff) | |
| download | musl-1e5eb73545ca6cfe8b918798835aaf6e07af5beb.tar.gz musl-1e5eb73545ca6cfe8b918798835aaf6e07af5beb.tar.xz musl-1e5eb73545ca6cfe8b918798835aaf6e07af5beb.zip | |
math: add fma TODO comments about the underflow issue
The underflow exception is not raised correctly in some cornercases (see previous fma commit), added comments with examples for fmaf, fmal and non-x86 fma. In fmaf store the result before returning so it has the correct precision when FLT_EVAL_METHOD!=0
Diffstat (limited to 'src/math/fmal.c')
| -rw-r--r-- | src/math/fmal.c | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/src/math/fmal.c b/src/math/fmal.c index ccbe434d..87e30fcf 100644 --- a/src/math/fmal.c +++ b/src/math/fmal.c @@ -262,6 +262,8 @@ long double fmal(long double x, long double y, long double z) /* * There is no need to worry about double rounding in directed * rounding modes. + * TODO: underflow is not raised correctly, example in downward rounding: + * fmal(0x1.0000000001p-16000L, 0x1.0000000001p-400L, -0x1p-16440L) */ fesetround(oround); adj = r.lo + xy.lo; |