RE: MPICH-GM 1.2.5..9.

I believe the issue deals partially with the interoperability of the FORTRAN
function call interface. (meaning: case, underscores, and implicit string
length argument placement) 
Do ifc/efc and g77 use the same conventions?  
I don't know if cc=gcc and fc=ifc works when compiling mpich. (I would have
to test)

Mike

-----Original Message-----
From: Greg Lindahl [mailto:[email protected]] 
Sent: Tuesday, February 25, 2003 3:07 PM
To: [email protected]
Subject: Re: [Myrinet] MPICH-GM 1.2.5..9.

On Tue, Feb 25, 2003 at 01:20:23PM -0700, Christopher D. Maestas wrote:

> I still maintain that on an end to end approach, the intel compiler suite
would
> give better performance than the gcc suite, especially when trying to take
> advantage of Itanium2 and XEON intructions sets.  So with that in mind,
there is
> still an advantage in compiling mpich-gm with the intel compilers over
gcc.

I thought you said you'd actually done it and measured the difference.

If there's no proven performance advantage to compiling the Myrinet
stuff with the Intel compiler, then why should Myricom work around
bugs in the Intel compiler?

greg

_______________________________________________
Myrinet mailing list
[email protected]
http://email.osc.edu/mailman/listinfo/myrinet