<div dir="ltr"><div dir="ltr">Thank you, Lain,<br><br>I have made modification of <span>dbcsr_index_operations.F</span> according to the trunk 13197, now cp2k 2.4 is working fine with ifort 14.0.0.<br><br>cp2k 2.4 MPI version popt + openmpi 1.6.5 + ifort 14.0.0 + mkl 11.1 <br><br>CC = icc<br>CPP =<br>FC = mpif90<br>LD = mpif90<br>AR = ar -r<br>DFLAGS = -D__INTEL -D__FFTSG -D__parallel -D__BLACS -D__SCALAPACK -D__FFTW3 -D__HAS_smm_dnn -D__LIBINT -D__LIBXC2 -D__ELPA<br>MKLROOT=/opt/intel/mkl<br>INTEL_INC = $(MKLROOT)/include/fftw<br>CPPFLAGS = -C -traditional $(DFLAGS) -I$(INTEL_INC)<br>FCFLAGS = $(DFLAGS) -I$(INTEL_INC) -O3 -xavx -heap-arrays 64 -funroll-loops -fpp -free<br>FCFLAGS2 = $(DFLAGS) -I$(INTEL_INC) -O1 -xavx -heap-arrays 64 -fpp -free<br>LDFLAGS = $(FCFLAGS) -I$(INTEL_INC)<br>LIBS = -Wl,--start-group $(MKLROOT)/lib/intel64/libmkl_<wbr>blacs_openmpi_lp64.a $(MKLROOT)/lib/intel64/libmkl_<wbr>lapack95_lp64.a $(MKLROOT)/lib/intel64/libmkl_<wbr>scalapack_lp64.a -L$(MKLROOT)/lib/intel64 -lmkl_intel_lp64 -lmkl_core -lmkl_sequential $(MKLROOT)/lib/intel64/libmkl_<wbr>intel_thread.a $(MKLROOT)/interfaces/fftw3xf/<wbr>libfftw3xf_intel.a -Wl,--end-group \<br>-L/home/sang/Software/cp2k-2.<wbr>4.0/tools/build_libsmm/lib -lsmm_dnn \<br>/opt/libint/lib/libderiv.a \<br>/opt/libint/lib/libint.a \<br>/opt/etsf/lib/libxc.a \<br>-lstdc++ -lm -lpthread -liomp5<br><br>OBJECTS_ARCHITECTURE = machine_intel.o<br><br><br><br>On Friday, October 25, 2013 2:58:54 AM UTC+8, IBethune wrote:<blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Sang,
<br>
<br>It's hard to say what the problem is without more information, specifically:
<br>
<br>* a copy of your arch file
<br>* what version of MKL and Intel MPI you are using
<br>* what input file you are running
<br>
<br>What I can say is that with ifort 14.0.0, mkl 11.1 and impi 4.1.1.036 we have been able to successfully regtest sopt, popt and ssmp builds. There are still some problems with psmp (MPI+OpenMP) that we are investigating.
<br>
<br>You will also need a recent trunk version of the code (13197 or later) to get a workaround for a regression in ifort 14.0.0
<br>
<br>Cheers
<br>
<br>- Iain
<br>
<br>--
<br>
<br>Iain Bethune
<br>Project Manager, EPCC
<br>
<br>Email: <a>ibe...@epcc.ed.ac.uk</a>
<br>Twitter: @IainBethune
<br>Web: <a href="http://www2.epcc.ed.ac.uk/~ibethuneTel/Fax" target="_blank">http://www2.epcc.ed.ac.uk/~<wbr>ibethune
<br>Tel/Fax</a>: +44 (0)131 650 5201/6555
<br>Mob: +44 (0)7598317015
<br>Addr: 2404 JCMB, The King's Buildings, Mayfield Road, Edinburgh, EH9 3JZ
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>On 24 Oct 2013, at 15:19, Hongqian Sang wrote:
<br>
<br>> Hi Iain,
<br>>
<br>> We are also using ifort 14.0.0 and saw the same error.
<br>>
<br>> Adding two lines at the end of arch file, cp2k is running now.
<br>> dbcsr_index_operations.o: dbcsr_index_operations.F
<br>> $(FC) -c $(FCFLAGS2) $<
<br>> The first SCF step gave right energy, then it went crazy and never converged.
<br>>
<br>>
<br>> Cheers,
<br>> Sang
<br>>
<br>> 在 2013年9月30日星期一UTC+8下午3时56分26秒,<wbr>IBethune写道:
<br>> Yes, we are using 13.1.3 from XE 2013 update 5 at the moment and it is working fine.
<br>>
<br>> Cheers
<br>>
<br>> - Iain
<br>>
<br>> --
<br>>
<br>> Iain Bethune
<br>> Project Manager, EPCC
<br>>
<br>> Email: <a>ibe...@epcc.ed.ac.uk</a>
<br>> Twitter: @IainBethune
<br>> Web: <a href="http://www2.epcc.ed.ac.uk/~ibethune" target="_blank">http://www2.epcc.ed.ac.uk/~<wbr>ibethune</a>
<br>> Tel/Fax: +44 (0)131 650 5201/6555
<br>> Mob: +44 (0)7598317015
<br>> Addr: 2404 JCMB, The King's Buildings, Mayfield Road, Edinburgh, EH9 3JZ
<br>>
<br>>
<br>>
<br>>
<br>>
<br>>
<br>>
<br>>
<br>> On 29 Sep 2013, at 06:24, Joe Greenstone wrote:
<br>>
<br>> > Hi Iain,
<br>> >
<br>> > Had also upgraded to ifort 14.0.0 in composer xe 2013 sp1.
<br>> >
<br>> > Are you now using ifort 13.1.3 in composer xe 2013 update 5?
<br>> >
<br>> >
<br>> >
<br>> > On Saturday, 28 September 2013 05:28:10 UTC+9:30, IBethune wrote:
<br>> > Hi Samuel. We have also just upgraded to the intel 14.0 compiler in compeer ex 2013 sp 1 and saw the same problem for various test cases. I've reported a bug to Intel and am currently waiting for them to investigate. For now it's better to use the 13.1 compiler!
<br>> >
<br>> > --
<br>> > You received this message because you are subscribed to the Google Groups "cp2k" group.
<br>> > To unsubscribe from this group and stop receiving emails from it, send an email to <a>cp2k+...@googlegroups.com</a>.
<br>> > To post to this group, send email to <a>cp...@googlegroups.com</a>.
<br>> > Visit this group at <a href="http://groups.google.com/group/cp2k" target="_blank">http://groups.google.com/<wbr>group/cp2k</a>.
<br>> > For more options, visit <a href="https://groups.google.com/groups/opt_out" target="_blank">https://groups.google.com/<wbr>groups/opt_out</a>.
<br>>
<br>>
<br>> --
<br>> The University of Edinburgh is a charitable body, registered in
<br>> Scotland, with registration number SC005336.
<br>>
<br>>
<br>> --
<br>> You received this message because you are subscribed to the Google Groups "cp2k" group.
<br>> To unsubscribe from this group and stop receiving emails from it, send an email to <a>cp2k+...@googlegroups.com</a>.
<br>> To post to this group, send email to <a>cp...@googlegroups.com</a>.
<br>> Visit this group at <a href="http://groups.google.com/group/cp2k" target="_blank">http://groups.google.com/<wbr>group/cp2k</a>.
<br>> For more options, visit <a href="https://groups.google.com/groups/opt_out" target="_blank">https://groups.google.com/<wbr>groups/opt_out</a>.
<br>
<br>
<br>--
<br>The University of Edinburgh is a charitable body, registered in
<br>Scotland, with registration number SC005336.
<br>
<br></blockquote></div></div>