Re: amber-developers: qmmm2/neb_gb;compile problem with evb_input.f (ifort on linux)

From: Robert Duke <rduke.email.unc.edu>
Date: Tue, 28 Mar 2006 06:05:28 -0700

Dave -
Thanks for the tip about -lscs on the altix. I think since I only use the

vector math library currently, I can get pmemd to build on all altix
platforms I have encountered to date. However, if problems start arising,

we can try this. I had problems with sander, and it had to do, if I
remember correctly, with stuff needed by some blas function found in the
MKL
or another; I just decided that the 9/7.21 combo was pressing the limits
for
intel compatibility, and did get closer versions built out at ncsa on
cobalt.
Regards - Bob

----- Original Message -----
From: "David A. Case" <case.scripps.edu>
To: <amber-developers.scripps.edu>
Sent: Tuesday, March 28, 2006 1:06 AM
Subject: Re: amber-developers: qmmm2/neb_gb;compile problem with
evb_input.f
(ifort on linux)


> On Tue, Mar 28, 2006, Robert Duke wrote:
>
>> Regarding ifort 9, the problem I had was when combining it with MKL 8
>> (latest release), and the problem is that MKL 8 and ifort 9 both ship
>> with
>> incompatible libguides. So if you pick up "the wrong one" (and either
>> one
>
> Sorry: I remember now your previous posts on this...it just slipped my
> mind.
>
>
>> I also encountered link problems when using ifort 9 with mkl 7.21 on an

>> altix
>> box at unc when trying to build sander.
>
> SGI provides a -lscs library that has always worked well for me. That
> should
> always take precedence over MKL, I think. I should probably put
something
> in the configure script to ignore MKL on sgi_altix. I don't know if
this
> would help with pmemd or not.
>
> ...regards...dave
>
>
Received on Wed Apr 05 2006 - 23:49:33 PDT
Custom Search