[Wien] error, end-of-file during read

[Gavin Abo]

If you use -DFFTW3, "mv fftw3.f03 fftw3.f03_orig" also in SRC_lapw2 and 
SRC_hf

If you don't use -DFFTW3, apply the fft_modules.patch and vresp.patch 
[http://zeus.theochem.tuwien.ac.at/pipermail/wien/2012-November/017911.html]

On 3/12/2013 7:49 AM, Laurence Marks wrote:
> This may not be easy to fix since I suspect that the file fftw3.f03 is
> setup for a 64 bit machine. For certain you have one problem:
>
> Put -I/opt/local/fftw3/include at the beginning of your compile
> options -- simplest is to edit the Makefile in SRC_lapw0. What you
> currently have will pickup the fftw3,h file from
> /opt/intel/Compiler/11.0/073/mkl/include/fftw rather than the one you
> compiled.
>
> If this does not work try doing "mv fftw3.f03 fftw3.f03_orig" within
> SRC_lapw0 as well. This will then pickup the fftw3.f03 file from
> /opt/local/fftw3/include.
>
> If neither of these work then I suspect that there is no easy fix and
> best is to just remove the -DFFTW3 completely. The fourier transforms
> are not the slowest part of the scf loop and you will not lose much.
> With a 32 bit machine you are going to be limited to very simple
> systems.
>
> On Tue, Mar 12, 2013 at 8:20 AM, mouhamed mahdi <mouh2009 at gmail.com> wrote:
>>> N.B. I assume that the "..." in your options is just because you shortened
>>> the line in the email. Also, do you really have a 32 bit machine?
>>   - yes, and yes
>> after adding  -DFFTW3 in FOPT and compiling i got this errors:
>>
>> ifort -w90 -FR -O3 -mp1 -w -prec_div -pad -align -DINTEL_VML -DFFTW3
>> -traceback -I/opt/intel/Compiler/11.1/073/mkl/lib/32
>> -I/opt/intel/Compiler/11.1/073/mkl/include -I/usr/local/include
>> -I/opt/intel/Compiler/11.0/073/mkl/include/fftw  -c fft_modules.F
>> ./fftw3.f03(1240): error #6684: This is an incorrect value for a kind type
>> parameter in this context.   [C_LONG_DOUBLE_COMPLEX]
>>        complex(C_LONG_DOUBLE_COMPLEX), dimension(*), intent(out) :: in
>> --------------^
>> ./fftw3.f03(1241): error #6684: This is an incorrect value for a kind type
>> parameter in this context.   [C_LONG_DOUBLE_COMPLEX]
>>        complex(C_LONG_DOUBLE_COMPLEX), dimension(*), intent(out) :: out
>>
>>
>
>