Boost logo

Ublas :

From: Gunter Winkler (guwi17_at_[hidden])
Date: 2007-03-12 08:45:40


On Monday 12 March 2007 13:35, Preben Hagh Strunge Holm wrote:

> Hi sorry, you may have misunderstood my question. I should have written
> which flag to make sure that the same floating point arithmetic is used
> on different architectures!

There is only -ffloat-store that gcc tries to be as close as possible at IEEE
standard. And you can define -fno-unsafe-math-optimizations.

quote from gcc doc:
       -ffast-math
           Sets -fno-math-errno, -funsafe-math-optimizations, -fno-trap‐
           ping-math, -ffinite-math-only, -fno-rounding-math, -fno-signal‐
           ing-nans and fcx-limited-range.

mfg
Gunter