On 1/22/2011 7:33 PM, Philip Potter wrote: > On 23 January 2011 03:55, Greg Lindahl <greg@blekko.com> wrote: >> On Sat, Jan 22, 2011 at 08:39:13AM -0800, Chip Salzenberg wrote: >>> After carefully considering http://blog.andreas.org/display?id=9 I believe >>> we should add the flag "-fexcess-precision=standard" when compiling on gcc >>> 4.5+. Any controversy here? >> That flag is the right thing to do -- I'm surprised that it wasn't >> noticed earlier. It's not hard to write a little test in C so that we >> would know if other compilers need a similar switch, too. > That, or -mfparch=sse where possible. It's a faster safe method. Nice. Any gcc version info on that one? Or should we just probe for it?Thread Previous | Thread Next