asm for hypot and hypotf
authorRich Felker <dalias@aerifal.cx>
Fri, 23 Mar 2012 05:52:49 +0000 (01:52 -0400)
committerRich Felker <dalias@aerifal.cx>
Fri, 23 Mar 2012 05:52:49 +0000 (01:52 -0400)
commitad2d2b963a4bf9e2631b345c898e8715b36b459e
treea99362612a0a46e2601aa081b80a9ca5cb1a5d83
parenta9e85c0a5c690086c652f3a8ae9a109389f671a3
asm for hypot and hypotf

special care is made to avoid any inexact computations when either arg
is zero (in which case the exact absolute value of the other arg
should be returned) and to support the special condition that
hypot(┬▒inf,nan) yields inf.

hypotl is not yet implemented since avoiding overflow is nontrivial.
src/math/i386/hypot.s [new file with mode: 0644]
src/math/i386/hypotf.s [new file with mode: 0644]