[Pw_forum] [Pw_users] Pre-release of Quantum-Espresso v.4

[Giovanni Cantele]

Axel Kohlmeyer wrote:
> On Wed, 5 Mar 2008, Giovanni Cantele wrote:
>
>
> dear giovanni,
>
> [...]
> GC> Dear all,
> GC> I have a problem in compiling the cvs versions of Quantum-ESPRESSO on an
> GC> alpha machine. The problem arises because the compilation of iotk, 
> GC> always working
> GC> up to 3.2.3, now stops at a routine that, as far as I understand, has 
> GC> been added after the
> GC> 3.2 release (but not included in 3.2.x patches):
> GC> 
> GC> f90 -tune ev7 -arch ev7 -fast -O4 -free -cpp -D__ALPHA -D__FFTW 
> GC> -D__USE_INTERNAL_FFTW -D__MPI -D__PARA -I../include  -I./  -I../Modules  
> GC> -I../iotk/src  -I../PW  -I../PH  -I../CPV -c iotk_tool.f90
> GC> 
> GC> a huge number of errors as the ones listed below follows:
> GC> 
> GC> f90: Error: iotk_tool.spp, line 477: Unrecognized token '#' skipped
> GC> # 475 "iotk_tool.spp"
> GC> ^
>
> this looks like a compiler bug. you can remove that specific line
> (it is just a comment). in fact, try the following:
>
> mv iotk/src/iotk_tool.f90 iotk/src/iotk_tool.f90.saved
> grep -v '^#' iotk/src/iotk_tool.f90.saved > iotk/src/iotk_tool.f90
>
> and then recompile.
>
> let us know if that helps.
>
> cheers,
>    axel.
>   
Dear Axel,
thanks a lot for reply.

Actually, that doesn't help.
After, I find errors like
f90: Error: iotk_tool.f90, line 418: This name does not have a type, and 
must have an explicit type.   [IOTK_LOGICAL2]
  write(*,"(a)") "  logical(kind="//trim(iotk_itoa(iotk_logical2))//")"
---------------------------------------------------^
f90: Error: iotk_tool.f90, line 418: There is no matching specific 
function for this generic function reference.   [IOTK_ITOA]
  write(*,"(a)") "  logical(kind="//trim(iotk_itoa(iotk_logical2))//")"
-----------------------------------------^

I'm not sure that the problems is with the lines beginning with '#' 
because most of the files within iotk/src are built with the same logic
(try: grep '^#' *f90), but their compilation succeeds. It seems like 
some "include" or some preprocessing instruction is not given to only
iotk_tool.f90, so as it fails. I inspected also the Makefile, to see if 
by chance some .o file was missing in the iotk_tool.o dependencies, but
I couldn't find one.
Actually, I couldn't well understand how *.spp files are used before the 
'true' compilation.

Giovanni


PS did someone succeed in compiling the CVS on alpha-Tru64 Unix?

> GC> f90: Error: iotk_tool.spp, line 477: Syntax error, found 
> GC> CHARACTER_CONSTANT 'iotk_tool.spp' when expecting one of: * :: , 
> GC> <END-OF-STATEMENT> ; : ) + . - (/ [ ] /) ' ** / > // .LT. ...
> GC> # 475 "iotk_tool.spp"
> GC> ------^
> GC> f90: Error: iotk_tool.spp, line 478: Syntax error, found 
> GC> END-OF-STATEMENT when expecting one of: (
> GC> " intro"
> GC> --------^
> GC> 
> GC> 
> GC> After I tried to "touch iotk_tool.o", the "make libiotk.a" ended with no 
> GC> error, therefore iotk_tool.f90 is the
> GC> only file which the compiler does not agree with.
>
>
> GC> 
> GC> Any hint about how to "make iotk_tool.o" ???
> GC> 
> GC> I'm compiling under OSF1 V5.1 2650 alpha
> GC> 
> GC> compiler version: HP Fortran Compiler V5.5A-3548-48D88
> GC> 
> GC> 
> GC> Thanks a lot,
> GC> 
> GC> Giovanni
> GC> 
> GC> 
>
>   


-- 



Dr. Giovanni Cantele
Coherentia CNR-INFM and Dipartimento di Scienze Fisiche
Universita' di Napoli "Federico II"
Complesso Universitario di Monte S. Angelo - Ed. 6
Via Cintia, I-80126, Napoli, Italy
Phone: +39 081 676910
Fax:   +39 081 676346
E-mail: giovanni.cantele at cnr.it
        giovanni.cantele at na.infn.it
Web: http://people.na.infn.it/~cantele
Research Group: http://www.nanomat.unina.it