Re: MacOSX 64bit binaries

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Wed Dec 03 2008 - 15:14:23 CST

please note also, that the g++ supplied by apple is
rumored to be "buggy" or "nonstandard" in many ways
and that may also contribute to many problems. one
major issue that may apply to all g++/gcc compiles is
that older code are not fully ANSI c/c++ compliant with
respect to aliasing. gcc version 4 assumes this by default
and this can easily lead to hard to track down miscompiles
(but also faster code for ANSI compliant sources).
so quite often, adding -fno-ansi-aliasing to compiles with
newer gcc, can work around these problems.

cheers,
    axel.

On 12/3/08, Giacomo Fiorin <gfiorin_at_seas.upenn.edu> wrote:
> Hello Chris, since it crashes in between there may be indeed something
> that g++ is not happy about in one of the NAMD files. The CVS source
> tree undergoes many changes, so this is likely to happen sometimes.
> Can you past the Darwin-XXX option that you used as an option to
> ./config, and the error message in the compilation?
>
>
>
> ---- -----
> Giacomo Fiorin
> Center for Molecular Modeling at
> University of Pennsylvania
> 231 S 34th Street, Philadelphia, PA 19104-6323
> phone: (+1)-215-573-4773
> fax: (+1)-215-573-6233
> mobile: (+1)-267-324-7676
> mail: giacomo.fiorin_<at>_gmail.com
> web: http://www.cmm.upenn.edu/
> ---- ----
>
>
>
> On Wed, Dec 3, 2008 at 2:34 AM, Christopher Hartshorn
> <cmhartshorn_at_wsu.edu> wrote:
>> Hello all. I am trying to build a set of 64bit binaries for 8core xeon
>> Mac
>> Pro's and have run into a problem. I was able to successfully build a the
>> Charm++ binaries in 64bit (I know they are because when I run the
>> megatests
>> they show up in Activity Monitor as 64bit Intel processes) fully optimized
>> for this machine (although only using gcc, icc keeps crashing because of a
>> problem with icpc?). But when I go to build NAMD (CVS) using a link to
>> the
>> charm folder (in the Mac-i686 arch file) which contains
>> "net-darwin-x86_64",
>> it crashes out on me midway or so through the compilation. To make sure I
>> am doing the right things, I tested my skills by doing the exact same
>> procedure using a set of 32 bit Charm++ binaries which I built, as well,
>> and
>> NAMD compiled completely with these. Also, are there any compiler flags
>> that can be set when compiling NAMD (like when one compiles Charm++) and
>> where would they be set? Thanks for any help.
>>
>> Regards,
>> Chris Hartshorn
>> WSU
>>
>>
>
>

-- 
=======================================================================
Axel Kohlmeyer   akohlmey_at_cmm.chem.upenn.edu   http://www.cmm.upenn.edu
  Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:48:44 CST