[SAC-HELP] intel mac and fortran

Georgia Cua georgia.cua at sed.ethz.ch
Sun Aug 20 09:50:47 PDT 2006


Dear Arthur and others who've posted on my question regarding intel  
mac, fortran, and sac,

Thanks very much for the recent posts on this topic.

It is very useful for me to know that sacio.a should not work on the  
intel mac. (As I wasn't sure whether the problems were with the  
compilers I was using.)

I will try sac version 100.1, and will write the list about whether  
it works or not on the intel mac.

I got my gfortran compiler from http://hpc.sourceforge.net.

The installation instructions are pretty straighforward (just  
download and untar).

It did not work for me immediately upon installation. At first, I got  
the following errors:

gfortran hello.f90
/usr/bin/ld: Undefined symbols:
___dso_handle
collect2: ld returned 1 exit status

I googled these errors, and tried updating my cctools, as suggested  
in the help listings on this site: http://gcc.gnu.org/ml/gcc-bugs/ 
2006-04/msg00895.html.

There is a link for cctools for darwin on: http://gcc.gnu.org/ml/gcc/ 
2006-03/msg00507.html

gfortran seemed to work for me after this (or at least, compiled a  
simple hello world program)

Thanks again,
Georgia



On Aug 19, 2006, at 6:53 PM, Arthur Snoke wrote:

> Georgia,
>
> Before I get to your question/situation, please share where you got  
> your gfortran compiler from.  I am told that this version is the  
> first one that has "acceptably fast" I/O.
>
> 1.  I am told that applications written for a PPC can run on the  
> Intel mac.  The translation is done using "Rosetta."  Applications  
> will probably run much slower, but they will run.  Hence you are  
> able to run the PPC compiled version of sac 100.0 from March 2005  
> on your intel mac.
>
> Version 100.1 that came out in July 2005 has command-line editing  
> using the libedit package.  Libedit has at least one dynamically  
> loaded library. It would be interesting to see if that version,  
> compiled on a PPC, could run on the intel mac.  That version can be  
> gotten from IRIS by the same procedure as you got 100.0.
>
> 2.  I do not think the emulation capabilities goes to the extent of  
> libraries, like sacio.a.  That is, sacio.a built on a PPC cannot be  
> used to build programs on an intel mac.  In addition to the endian  
> difference, one has to use a Fortran compiler configured with a gcc  
> 3 compiler, as that binary version of sac 100.0 was built using gcc  
> 3.  George Hellfrich sent a link to a URL that should allow you to  
> build a g77 that is configured with gcc 3, but that will not help  
> you until you get a sacio.a library built on an intel mac.
>
> 3.  An earlier posting reported failure in an attempt to build sac  
> from sources on an intel mac.  Researchers at IRIS are looking into  
> this.
>
> 4.  I have a PPC with 10.4.7.  I am using gcc 3.3 and g77  
> configured with gcc 3.4.6.  I am told that g77 will not be  
> supported in the future so that one will eventually need to go to  
> gfortran.  Hence my question above about the source of gfortran  
> configured with gcc 4.2 (experimental).  I think there is no  
> getting around the fact that one needs a sacio.a built with the  
> same level of gcc (3 or 4) as your Fortran compiler -- and with the  
> same endian.
>
> Not sure how much help this is.  If you or anyone else see flaws  
> with what I have written, please share.
>
> Arthur



-----------------------------
Dr. Georgia Cua
Swiss Seismological Service
Institute of Geophysics
ETH Hoenggerberg HPP P
CH-8093 Zurich
Switzerland

Tel:  +41-044-633-7574
Fax: +41-044-633-1065

georgia.cua at sed.ethz.ch
-----------------------------


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.iris.washington.edu/pipermail/sac-help/attachments/20060820/ce5ec5be/attachment.html>


More information about the sac-help mailing list