Mail Archives: djgpp-workers/2003/12/16/12:25:28
--part1_59.17546f3.2d109998_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Kudos to the person who put a note in a patch for e_pow.c concerning cases
where, for z=pow(x,y), when x is very near 1 and y is very large, z would
vanish. It prompted me to create test cases for all three floating point
precisions. The same problem existed for the float version and was easily fixed.
For testing the double precision case, the parameters are:
x = 1-DBL_EPSILON
y = log(DBL_MIN)/log(1-DBL_EPSILON)
KB Williams
--part1_59.17546f3.2d109998_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
<HTML><FONT FACE=3Darial,helvetica><HTML><FONT SIZE=3D3 PTSIZE=3D12 FAMILY=
=3D"SERIF" FACE=3D"Georgia" LANG=3D"0">Kudos to the person who put a note in=
a patch for e_pow.c concerning cases where, for z=3Dpow(x,y), when x is ver=
y near 1 and y is very large, z would vanish. It prompted me to create=
test cases for all three floating point precisions. The same problem=20=
existed for the float version and was easily fixed.<BR>
<BR>
For testing the double precision case, the parameters are:<BR>
<BR>
x =3D 1-DBL_EPSILON<BR>
y =3D log(DBL_MIN)/log(1-DBL_EPSILON)<BR>
<BR>
<BR>
KB Williams</FONT></HTML>
--part1_59.17546f3.2d109998_boundary--
- Raw text -