[Bug math/14469] Inaccurate j0f function

[zimmerma+gcc at loria dot fr]

https://sourceware.org/bugzilla/show_bug.cgi?id=14469

--- Comment #5 from Paul Zimmermann <zimmerma+gcc at loria dot fr> ---
the large error for huge operands can be avoided by removing the
optimization "if(ix>0x48000000)" around line 63 of
sysdeps/ieee754/flt-32/e_j0f.c. Then for x=0x1.4665d2p24 (x=21390802) we get an
error of 239 ulps
only (which corresponds to the cancellation between u*cc and v*ss in the
expression u*cc-v*ss at line 66).

-- 
You are receiving this mail because:
You are on the CC list for the bug.