Page 1 of 2
Compiling on IBM power6 in CINECA
Posted: Fri Oct 09, 2009 4:47 pm
by feffeficus
Hi all. sorry to bother, but do someone of you have problems in compiling yambo on
sp6 in cineca?
I load this modules:
Currently Loaded Modulefiles:
1) profile/base 4) fftw/3.2.2--xl--10.1
2) QuantumESPRESSO/4.1 5) netcdf/4.0.1--xl--10.1
3) xl/10.1
I compile this yambo:
yambo-3.2.1-rev486
like that configure using:
Code: Select all
./configure
--with-fftw=/cineca/prod/libraries/fftw/3.2.2/xl--10.1/
--with-netcdf-lib=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/lib/
--with-netcdf-include=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/include/
--with-p2y=4.0 --with-iotk=/sp6/userhpe/hpefrueh/codes/espresso-4.1/iotk
--build=powerpc-ibm
CC=cc
CPP=cpp
FCFLAGS="-O2 -q64 -qstrict -qarch=pwr6 -qtune=pwr6 -qmaxmem=-1 -qsuffix=f -qsuppress=1501-245 *.c"
and i get this error message:
checking for Fortran flag to compile .f files... unknown
configure: error: Fortran could not compile .f files
Where is the problem?
Re: Compiling on IBM power6 in CINECA
Posted: Sat Oct 10, 2009 2:49 am
by Daniele Varsano
Ciao Feffe,
I compiled the code in sp6 using:
module load xl/10.1
module load netcdf/4.0.1--xl--10.1
module load fftw/3.2.2--xl--10.1
module load QuantumESPRESSO/4.1
export CPP=cpp
export CC=xlc_r
export F77=xlf_r
export FC=xlf90_r
export FCFLAGS='-O2 -q64 -qstrict -qarch=pwr6 -qtune=pwr6 -qmaxmem=-1 -qsuffix=f=f'
./configure --with-fftw=/cineca/prod/libraries/fftw/3.2.2/xl--10.1/ --with-netcdf-lib=/cineca/prod/librari
es/netcdf/4.0.1/xl--10.1/lib --with-netcdf-include=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/include --w
ith-p2y=4.0 --with-iotk=/cineca/prod/build/applications/QuantumESPRESSO/4.1/xl--10.1/BA_WORK/QuantumESPRESS
O-4.1/iotk --build=powerpc-ibm
it should work,
Daniele
Re: Compiling on IBM power6 in CINECA
Posted: Mon Oct 12, 2009 12:32 pm
by feffeficus
perfect!!
it compiles!
my fault: after the ./configure step i used make instead of gmake.
now with gmake it's ok!
many thanks again!
ff
Re: Compiling on IBM power6 in CINECA
Posted: Thu Mar 18, 2010 1:58 pm
by Letizia Chiodo
Dear developers,
I'm trying to compile yambo, version 3.2.1-r.448, on cineca sp6.
I followed the suggestions given here in this discussion, and (after gmake clean_all)
also the options given in another discussion of the forum,
on the same topic (it's the one titled "Configure Error on IBM Power6").
but in both cases I cannot compile yambo, the error is the same, in the routine tddft_do_X_W_typs.f.,
attached you can find the error in compiling, while I couldn't attach the config.log because too big.
any kind of help/suggestion is appreciated
thanks a lot
letizia
Letizia Chiodo
IIT- Italian Institute of Technology, and ETSF
Re: Compiling on IBM power6 in CINECA
Posted: Thu Mar 18, 2010 2:08 pm
by andrea marini
Try removing all *.f files. If any files named *.f is present then the compilation will fail. In the next distro Makefile will be chanegd in order to clean these intermediate preprocessed files.
Re: Compiling on IBM power6 in CINECA
Posted: Fri Mar 19, 2010 4:33 pm
by Letizia Chiodo
Thanks Andrea,
I could compile by restarting from the .tar.gz,
by using
./configure --build=powerpc-ibm --with-fftw=/cineca/prod/libraries/fftw/3.2.2/xl--10.1/lib --with-netcdf-lib=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/lib --with-netcdf-include=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/include --with-iotk=/cineca/prod/build/applications/QuantumESPRESSO/4.1/xl--10.1/BA_WORK/QuantumESPRESSO-4.1/iotk --with-p2y=4.0
however, yambo works, but p2y doesn't.
it produces only the ns.db1, not the ns.wf.
ns.db1 content looks fine with yambo -D
RD./SAVE//ns.db1]------------------------------------------
Bands : 20
K-points : 1
G-vectors [RL space]: 922329
Components [wavefunctions]: 115313
Symmetries [spatial+T-rev]: 16
Spinor components : 1
Spin polarizations : 1
Temperature [ev]: 0.000000
Electrons : 18.00000
WF G-vectors : 115313
Max atoms/species : 1
No. of atom species : 2
- S/N 008688 ---------------------------- v.03.02.01 r.448 -
but p2y stops when reading wavefunctions, with the error
ERROR: 0031-250 task 0: Trace/BPT trap
the problem is not in the espresso database, because I tried with a different system, and also tried both espresso 4.0.5 and espresso 4.1 (the sp6 preinstalled version).
always the same error.
thanks
ciao
letizia
Re: Compiling on IBM power6 in CINECA
Posted: Fri Mar 19, 2010 6:48 pm
by Conor Hogan
Hi Lety,
I've been using p2y compiled with those flags for my systems, and its been working ok with the following script.
Code: Select all
#!/bin/bash
#
# @ job_name = myjob
# @ output = $(jobid).out
# @ error = $(jobid).err
# @ wall_clock_limit = 0:10:00
# @ total_tasks = 1
# @ task_affinity=cpu(1)
# @ job_type = parallel
# @ resources = ConsumableMemory(920Mb)
#
# @ queue
../yambo/branches/gpl_sp6/bin/p2y -S
Can you see if its just a problem with the queue and the memory? How were you submitting the job?
Does a much smaller system work ok? Your system looks pretty large...maybe the -S flag can help.
Re: Compiling on IBM power6 in CINECA
Posted: Sun Mar 21, 2010 2:58 pm
by Conor Hogan
Dear Letizia,
Using yambo/p2y compiled with:
Code: Select all
./configure --build=powerpc-ibm --with-fftw=/cineca/prod/libraries/fftw/3.2.2/xl
--10.1/lib --with-netcdf-lib=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/lib --
with-netcdf-include=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/include --with-
iotk=/cineca/prod/build/applications/QuantumESPRESSO/4.1/xl--10.1/BA_WORK/Quantu
mESPRESSO-4.1/iotk --with-p2y=4.0
submitting the job with
Code: Select all
#!/bin/bash
#
# @ job_name = myjob
# @ output = $(jobid).out
# @ error = $(jobid).err
# @ wall_clock_limit = 0:10:00
# @ total_tasks = 1
# @ task_affinity=cpu(1)
# @ job_type = parallel
# @ resources = ConsumableMemory(920Mb)
#
# @ queue
../yambo/branches/gpl_3.2.1_r448/bin/p2y -v
with the data in ZnO-ns-4.1.save,
I get a successful generation of SAVE:
Code: Select all
<---> Energies...
<---> :: Electrons : 18.00000
<---> :: Temperature [ev]: 0.000000
<---> :: Lattice factors [a.u.]: 20.00000 20.00000 20.00000
<---> :: K-points : 1
<---> :: Bands : 20
<---> :: Spinor components : 1
<---> :: Spin polarizations : 1
<---> :: Spin orbit coupling : no
<---> :: Symmetries [spatial]: 8
<---> :: [T-rev]: yes
<---> :: Max WF components : 115313
<---> :: RL vectors (WF): 115313
<---> :: RL vectors (CHARGE): 922329
<---> :: XC potential : Perdew & Wang (xc)
<---> :: Atomic species : 2
<---> :: Max atoms/species : 1
<---> == Writing DB1 ...
<---> == DB2 (wavefunctions) ...
<---> [M 0.018 Gb] Alloc wf_disk (0.018)
<---> :: K-point: 1/ 1 Spinors 1
Code: Select all
sp0201 38% ls -lrt SAVE
total 25856
18450672 Mar 21 14:54 ns.wf
11530940 Mar 21 14:54 ns.db1
as confirmed by yambo -D:
Code: Select all
[RD./SAVE//ns.db1]------------------------------------------
Bands : 20
K-points : 1
G-vectors [RL space]: 922329
Components [wavefunctions]: 115313
Symmetries [spatial+T-rev]: 16
Spinor components : 1
Spin polarizations : 1
Temperature [ev]: 0.000000
Electrons : 18.00000
WF G-vectors : 115313
Max atoms/species : 1
No. of atom species : 2
- S/N 006598 ---------------------------- v.03.02.01 r.448 -
[M 0.012 Gb] Alloc g_vec k_pt dl_sop wf_nc_k wf_igk (0.012)
[RD./SAVE//ns.wf]-------------------------------------------
Bands in each block : 20
Blocks : 1
- S/N 006598 ---------------------------- v.03.02.01 r.448 -
so...no idea why you are having that problem! If you solve it, please let us know.
Conor
Re: Compiling on IBM power6 in CINECA
Posted: Wed Mar 24, 2010 6:44 pm
by Letizia Chiodo
Hi Conor,
finally I solved, but don't ask me how.
I simply restarted from .tar.gz,
recompiled the interfaces with the configure you suggested,
and it worked.
so, thanks for your help!
cheers
letizia
Re: Compiling on IBM power6 in CINECA
Posted: Thu Apr 08, 2010 3:53 pm
by feffeficus
Hi guys!
Sorry again, but today I woke up and tried to compile yambo on SP6 at Cineca and I have 2 problems
1) last revision (yambo_rev 576)
with that configuration option:
Code: Select all
Currently Loaded Modulefiles:
1) profile/base 4) fftw/3.2.2--xl--10.1
2) QuantumESPRESSO/4.1 5) netcdf/4.0.1--xl--10.1
3) xl/10.1 6) gnuplot/3.7.1-ibm
and the option indicated here in Daniele's post:
Code: Select all
export CPP=cpp
export CC=xlc_r
export F77=xlf_r
export FC=xlf90_r
export FCFLAGS='-O2 -q64 -qstrict -qarch=pwr6 -qtune=pwr6 -qmaxmem=-1 -qsuffix=f=f'
./configure --build=powerpc-ibm --with-fftw=/cineca/prod/libraries/fftw/3.2.2/xl--10.1/lib --with-netcdf-lib=/cineca/prod/
libraries/netcdf/4.0.1/xl--10.1/lib --with-netcdf-include=/cineca/prod/libraries/netcdf/4.0.1/xl--10.1/include --with-iotk=/ci
neca/prod/build/applications/QuantumESPRESSO/4.1/xl--10.1/BA_WORK/QuantumESPRESSO-4.1/iotk --with-p2y=4.0
then when doing
gmake interfaces
I found this error:
Code: Select all
1501-511 Compilation failed for file qexml_v4.0.f.
gmake[1]: *** [qexml_v4.0.o] Error 1
gmake[1]: Leaving directory `/sp6/userhpe/hpefrueh/codes/yambo-3.2.1_rev576/interfaces/p2y'
gmake: *** [interfaces] Error 2
2) rev 485
I tried then with this old revision, with the same configuration above. It compiles both yambo and interfaces, but when doing initialization with
it writes only that in the yambo.in:
#
# Y88b / e e e 888~~\ ,88~-_
# Y88b / d8b d8b d8b 888 | d888 \
# Y88b/ /Y88b d888bdY88b 888 _/ 88888 |
# Y8Y / Y88b / Y88Y Y888b 888 \ 88888 |
# Y /____Y88b / YY Y888b 888 | Y888 /
# / / Y88b / Y888b 888__/ `88_-~
#
# GPL Version 3.2.1 Revision 485
#
http://www.yambo-code.org
#
setup # [R INI] Initialization
3) I tried then with another old revision 448 compiled by Letizia on Sp6 in the same way and it works properly.
Any idea?
(I can't attach the two confi.log files, too big. Ask if you wnat them..) sorry
Thanks in advance!
ff