bde 30aa45f24b Fix the hi+lo decomposition for 2/(3ln2). The decomposition needs to ...

be into 12+24 bits of precision for extra-precision multiplication,
but was into 13+24 bits.  On i386 with -O1 the bug was hidden by
accidental extra precision, but on amd64, in 2^32 trials the bug
caused about 200000 errors of more than 1 ulp, with a maximum error
of about 80 ulps.  Now the maximum error in 2^32 trials on amd64
is 0.8573 ulps.  It is still 0.8316 ulps on i386 with -O1.

The nearby decomposition of 1/ln2 and the decomposition of 2/(3ln2) in
the double precision version seem to be sub-optimal but not broken.

2008-02-14 10:23:51 +00:00

..

amd64

Use hardware remainder on amd64 since it is 5 to 10 times faster than

2008-02-13 06:01:48 +00:00

arm

Use C comments since we now preprocess these files with CPP.

2007-04-29 14:05:22 +00:00

bsdsrc

Fix tgamma() on some special args:

2007-05-02 15:24:49 +00:00

i387

Implement rintl(), nearbyintl(), lrintl(), and llrintl().

2008-01-14 02:12:07 +00:00

ia64

Use C comments since we now preprocess these files with CPP.

2007-04-29 14:05:22 +00:00

ld80

Fix exp2*(x) on signaling NaNs by returning x+x as usual.

2008-02-13 10:44:44 +00:00

ld128

Fix exp2*(x) on signaling NaNs by returning x+x as usual.

2008-02-13 10:44:44 +00:00

man

Introduce a new log(3) manpage and move the relevant functions there.

2008-01-18 21:43:00 +00:00

powerpc

Use C comments since we now preprocess these files with CPP.

2007-04-29 14:05:22 +00:00

sparc64

Use C comments since we now preprocess these files with CPP.

2007-04-29 14:05:22 +00:00

src

Fix the hi+lo decomposition for 2/(3ln2). The decomposition needs to

2008-02-14 10:23:51 +00:00

Makefile

On arches where long double is the same as double, alias ceil(), floor()

2008-02-13 16:56:52 +00:00

Symbol.map

Hook up exp2l() and related docs to the build.

2008-01-18 21:43:10 +00:00