From: John Maddock (john_at_[hidden])
Date: 2008-04-03 12:19:25
> Beman Dawes <bdawes <at> acm.org> writes:
>> Changes I'd like to see include:
>> * Drop the Intel compiler on Windows. Not enough people seem to care
>> about this compiler/platform.
> Whaa?? I would suggest that ICL 10.1 should be supported. It is
> probably the most conformant (and the best optimizing) compiler out
> there. Irritatingly, 10.0 had introduced many regressions over 9.1,
> but with 10.1 things appear to be back on track.
> Mind you, to get the best optimizations, some esoteric flags have to
> be used; also, by default the optimizer is set to favour speed at the
> cost of floating point accuracy. I have played around a bit with
> optimization settings, and can post the options I use to build boost
> components (since bjam does not include aggressive optimization
> settings), if someone would find it useful.
> Anyhow, fwitw, I strongly favour keeping ICL support - even if it is
> not used much owing to not being free, it would function an examplar
> to other vendors.
I'm inclined to agree to that: Intel have been extremely responsive in the
past in responding to compiler issues relating to Boost, and given the level
of conformance offered by this compiler, does it really soak up that much
time to officially support it? Incidently, it's good to have at least one
EDG based front end in our official support list, even if it does emulate
the MSVC bugs ;-)
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk