Thanks, Jason.
The source file edit did the trick.
Sasha
Jason Swails wrote:
> Hi Sasha,
>
> This has actually come up several times in the past few weeks (several == 2,
> IIRC). The issue here is that whichever version of gfortran you're using
> doesn't like the private, memTracker type to be "unsaved". That is, find
> this line in safemem.f (NOT _safemem.f, so it won't be line 107), and add
> the save attribute to that line (add ,save after private).
>
> This should be all that's necessary to fix it (is this worth a bugfix?)
>
> HTH,
> Jason
>
> On Wed, Oct 19, 2011 at 7:21 PM, Sasha Buzko <obuzko.ucla.edu> wrote:
>
>
>> Hi all,
>> Installing AmberTools on a CentOS 5.5 (x86_64) using gcc. Compilation of
>> AmberTools 1.5 produces the following error:
>>
>> cpp -traditional -P -DMKL -DRISM -DRISM_CRDINTERP -DRISM_DX
>> -DRISM_LINPROJ safemem.f > _safemem.f
>> gfortran -c -O3 -ffree-form -o safemem.o _safemem.f
>> In file _safemem.f:107
>>
>> type(memTracker),private :: totalMem
>> 1
>> Error: Object 'totalmem' at (1) must have the SAVE attribute for default
>> initialization of a component
>> make[1]: *** [safemem.o] Error 1
>> make[1]: Leaving directory `/data/amber11/AmberTools/src/rism'
>> make: *** [serial] Error 2
>>
>> Never had any issues with compilation of the previous versions of
>> AmberTools.
>>
>> Thanks for any suggestions.
>>
>> Sasha
>>
>>
>>
>>
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>>
>
>
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Oct 19 2011 - 18:30:03 PDT
