Re: [AMBER-Developers] error compiling amber

From: Jason Swails <>
Date: Tue, 2 Mar 2010 15:18:15 -0500

On Tue, Mar 2, 2010 at 2:27 PM, case <> wrote:
> On Tue, Mar 02, 2010, Jason Swails wrote:
>> I just checked out a version of amber earlier this morning and ran
>> into this error:
>> ifort -c -O0 -FR  -o xray_utils.o _xray_utils.f
>> _xray_utils.f(256): error #6423: This name has already been used as an
>> external function name.   [ERF_INV]
>>       erf_inv = a(1)
>> ------^
>> compilation aborted for _xray_utils.f (code 1)
> For now, you should just undo r1.5 for xray_utils, i.e. go back to 1.4
> (this is essentially the same as changing MKL to MKL_something else).
>> I'm using intel compilers and MKL all version 11.  I don't know if
>> it's an older version of MKL that doesn't have erf_inv?
> The error is at the compiliation step, not the link step, so it has nothing
> to do with what is or is not in the MKL libraries.
> Try changing line 214 of xray_utils.f from "function erf_inv(x)" to
> "pure function erf_inv(x)".  This is a guess, but see if it helps.

Gives more errors (including the original one), mainly complaining
that a pure function will only play with other pure functions. I'm
wary of changing these to pure (in case they're not), since each of
those functions may in turn require me to change more. It's far
easier to close my eyes and pretend that ifdef doesn't exist :). My
idea is that if erf_inv already exists, then it's doing the right
thing? I'll report back if tests fail.


> ...dac
> _______________________________________________
> AMBER-Developers mailing list

Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
AMBER-Developers mailing list
Received on Tue Mar 02 2010 - 12:30:03 PST
Custom Search