[Pw_forum] iotk error - again

Lorenzo Paulatto Lorenzo.Paulatto at impmc.upmc.fr
Thu Feb 10 12:08:58 CET 2011


Are you using the most recent version of intel fortran compiler (it's  
called XE, or 11.0)? It has a lot of problem, we've been discussing the  
issue some time ago: it definitely is a miscompilation problem. I have a  
stinky workaround, but the best solution is to downgrade the compiler.

cheers

On Thu, 10 Feb 2011 11:10:56 +0100, O. Baris Malcioglu  
<baris.malcioglu at gmail.com> wrote:

> Dear all,
> I've also started getting a lot of IOTK errors recently using the CVS
> version.
> I've got two possible candidates,
> 1)highly asynchronous file system (i.e. a parallel system that is not  
> robust
> in the sense of NFS, and highly overloaded)
> 2) a conspiracy going on with the newest intel compilers running in an
> environment that is not-so-new.
> Since this error is annoyingly frequent but not replicable at will, could
> you please contact me if you are using a asynchronous file system with  
> the
> compilers you are using? ,
> Best,
> Baris Malcioglu
>  P.S. The parallel file systems optimized for performance tend to be
> asynchronous, full list can be found here:
> http://en.wikipedia.org/wiki/List_of_file_systems
> <http://en.wikipedia.org/wiki/List_of_file_systems>
>
> ################################################################################
> ########################################
> # FROM IOTK LIBRARY, VERSION 1.2.0
> # UNRECOVERABLE ERROR (ierr=1)
> # ERROR IN: iotk_scan_tag (iotk_scan.spp:468)
> # CVS Revision: 1.23
> #
> iostat=67
> # ERROR IN: iotk_scan (iotk_scan.spp:743)
> # CVS Revision: 1.23
> # ERROR IN: iotk_scan_end (iotk_scan.spp:233)
> # CVS Revision: 1.23
> # foundl
> # ERROR IN: iotk_close_read (iotk_files.spp:645)
> # CVS Revision: 1.20
> ################################################################################
> ########################################
>
>
> 2010/12/8 W2AGZ <w2agz at w2agz.com>
>
>> Thanks, Paolo.  I'll look into it ths week.  -Paul
>>
>> -----Original Message-----
>> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
>> Behalf Of Paolo Giannozzi
>> Sent: Wednesday, December 08, 2010 1:49 AM
>> To: PWSCF Forum
>> Subject: Re: [Pw_forum] iotk error - again
>>
>> On Dec 7, 2010, at 8:33 , W2AGZ wrote:
>>
>> > "replace q_real by q_real_space."  Is this the fix?
>> >
>>
>>
>> no. This is the fix for a very specific case, and the only one I  
>> remember
>> in
>> which an error was issued by iotk for a good reason (i.e. a variable was
>> written with one name and read with another). All other instances of  
>> iotk
>> errors I remember are related to miscompilation.
>>
>> Try first of all to compile iotk with reduced optimization (e.g. using
>> $(MAKE) FFLAGS="$(FFLAGS_NOOPT)" )
>>
>> P.
>> ---
>> Paolo Giannozzi, Dept of Chemistry&Physics, Univ. Udine via delle  
>> Scienze
>> 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
>>
>>
>>
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://www.democritos.it/mailman/listinfo/pw_forum
>>
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://www.democritos.it/mailman/listinfo/pw_forum
>>


-- 
Lorenzo Paulatto (IdR)
IMPMC - CNRS UMR 7590 & Université P&M Curie
T23-C13/23-5e27 - 4 place Jussieu - 75252 Paris Cedex5
phone: +33 (0)144 27 5211
www:   http://www-int.impmc.upmc.fr/~paulatto/


More information about the Pw_forum mailing list