Closed rshopa closed 5 years ago
do you have mkl installed? if yes, make sure to
$> source `which compilervars.sh` intel64
before running cmake (assuming you're on a 64bit architecture). that should then fix the fftw3.h dependency.
c.
As far as I know, MKL is not installed.
ok, that explains why your issues with fftw3. i have never installed without mkl, so can't really tell much.
for now, could you deactivate the recon cmake target? there's another dependency on fftw3 (in wpg/srw), but that is handled by the wpg installation makefile and should work even without mkl.
carsten
----- Original Message ----- From: "Roman Shopa" notifications@github.com To: "eucall-software/simex_platform" simex_platform@noreply.github.com Cc: "Carsten Fortmann-Grote" carsten.grote@xfel.eu, "Comment" comment@noreply.github.com Sent: Tuesday, 26 March, 2019 14:52:04 Subject: Re: [eucall-software/simex_platform] SIMEX installation on cluster for Polish XFEL users (#170)
As far as I know, MKL is not installed.
-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/eucall-software/simex_platform/issues/170#issuecomment-476649420 -- Regards, Carsten
Dr. Carsten Fortmann-Grote Scientist for Scientific Simulations / Wissenschaftler fuer wissenschaftliche Simulationen European XFEL GmbH Holzkoppel 4 22869 Schenefeld Germany
Phone: +49 (0)40 8998-5603 Fax: +49 (0)40 8998-1905 Email: carsten.grote@xfel.eu Web: www.xfel.eu
Managing Directors: Prof. Dr. Robert Feidenhans'l
Registered as European X-Ray Free-Electron Laser Facility GmbH at Amtsgericht Hamburg, HRB 111165
UPD! Thanks a lot.
I've already added the whole module intel-composer
(with MKL) and, finally, CUDA (GAPDPhotonDiffractor=ON
). The complete list of modules uploaded (some might not be necessary) is as below:
(xfel) [rshopa@wn25048 build]$ module list
Currently Loaded Modulefiles:
1) cis-env 5) gcc/4.6.3 9) hdf5/1.8.17.patch1
2) cis-auto-remove 6) blas/3.5.0-x86_64-gcc46 10) bzip2/1.0.6
3) cuda/4.2.9 7) lapack/3.5.0-x86_64-gcc46 11) boost/1.64.0-x86_64-gcc71
4) binutils/2.22 8) fftw/3.3.4-x86_64-gcc482 12) intel-composer/2016.2.181
At first, errors appeared while make, but this was caused by assigning root dir to empty.
I changed manually INSTALL_PREFIX=/mnt/home/rshopa/SIMEX/simex_platform/
and the installation finally worked out.
BTW, CMake returns a warning -- all these vars were set empty as INSTALL_PREFIX at first. I hope it's not a problem.
Manually-specified variables were not used by the project:
BOOST_ROOT
FFTW_INCLUDES
FFTW_LIBRARIES
GEANT4_ROOT
XCSIT_ROOT
XERCESC_ROOT
There's issues on executing Test.py from unittest. Not sure it's an installation issue, but at least a warning of Esther
not being installed is odd
(xfel) [rshopa@wn25048 unittest]$ python Test.py -v 2>&1 | tee Test.log
...
WARNING: esther_execute could not be imported. This is most probably due to Esther not being installed or not found. Expect RunTimeErrors when attempting to run the EstherPhotonMatterInteractor.backengine().
WARNING: Importing libpy_detector_interface failed. This is most probably due to XCSIT and/or Geant4 not being installed properly on this system. The XCSITPhotonDetector class can still be instantiated, but the backengine() method will throw an exception.
WARNING: Importing libpy_detector_interface failed. This is most probably due to XCSIT and/or Geant4 not being installed or not being found. Expect errors.
There are other errors as follows:
...
testBackengineWithPropInput (SimExTest.Calculators.CrystFELPhotonDiffractorTest.CrystFELPhotonDiffractorTest)
Check that beam parameters can be taken from a given propagation output file. ... ERROR
testConstructionWithPropInput (SimExTest.Calculators.CrystFELPhotonDiffractorTest.CrystFELPhotonDiffractorTest)
Check that beam parameters can be taken from a given propagation output file. ... ERROR
...
testBackengine (SimExTest.Calculators.S2EReconstructionTest.S2EReconstructionTest)
Test that we can start a test calculation. ... ERROR
testBackengineDefaultPath (SimExTest.Calculators.S2EReconstructionTest.S2EReconstructionTest)
Test that we can start a test calculation. ... ERROR
testDefaultConstruction (SimExTest.Calculators.S2EReconstructionTest.S2EReconstructionTest)
Testing the default construction of the class. ... ok
testShapedConstruction (SimExTest.Calculators.S2EReconstructionTest.S2EReconstructionTest)
Testing the default construction of the class. ... ok
testBackengine (SimExTest.Calculators.SingFELPhotonDiffractorTest.SingFELPhotonDiffractorTest)
Test that we can start a test calculation. ... Traceback (most recent call last):
File "/mnt/home/rshopa/SIMEX/simex_platform/bin/radiationDamageMPI", line 9, in <module>
main(parameters=parameters)
File "/mnt/home/rshopa/SIMEX/simex_platform/lib/python3.6/pysingfel/radiationDamageMPI.py", line 26, in main
slave_diffract(comm, parameters)
File "/mnt/home/rshopa/SIMEX/simex_platform/lib/python3.6/pysingfel/radiationDamageMPI.py", line 94, in slave_diffract
MakeOneDiffr(myQuaternions, counter, parameters, outputName)
File "/mnt/home/rshopa/SIMEX/simex_platform/lib/python3.6/pysingfel/radiationDamage.py", line 160, in MakeOneDiffr
particle = Particle(inputName, datasetname)
File "/mnt/home/rshopa/SIMEX/simex_platform/lib/python3.6/pysingfel/particle.py", line 38, in __init__
self.readh5File(fname[0], fname[1])
File "/mnt/home/rshopa/SIMEX/simex_platform/lib/python3.6/pysingfel/particle.py", line 56, in readh5File
with h5py.File(fname, 'r') as f:
File "/mnt/home/rshopa/miniconda3/envs/xfel/lib/python3.6/site-packages/h5py/_hl/files.py", line 394, in __init__
swmr=swmr)
File "/mnt/home/rshopa/miniconda3/envs/xfel/lib/python3.6/site-packages/h5py/_hl/files.py", line 170, in make_fid
fid = h5f.open(name, flags, fapl=fapl)
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "h5py/h5f.pyx", line 85, in h5py.h5f.open
OSError: Unable to open file (file signature not found)
The program freezes at this moment (liable only to be broken) - is this just about long simulation time or due to errors?
can you make sure that Tests/python/unittest/TestFiles/pmi_out_0000001.h5 exists?
if not, you're probably missing git-lfs. please check https://help.github.com/en/articles/versioning-large-files.
c
[UPD!] Greetings.
I shall report problems with the basic installation of SIMEX platform on CIŚ cluster (Świerk, Poland https://www.cis.gov.pl/) for Polish XFEL users. Since docker is not available (Singularity is announced), it ought to be made manually as described on https://eucall-software.github.io/simex_platform/index.html.
I encountered multiple problems related with dependencies and links to the libraries. Apparently SIMEX requires some with Geant4 and CUDA which is not mentioned in docs. I turned off xcsit as Carsten advised and ran
install.sh
with the following optionsWhat exactly the function of
GAPDPhotonDiffractor
is appears a bit unclear to me and it was also turned off, since otherwise a CUDA-related error would erupt:I also had to set manually paths to FFTW3 library, however, an error occurs while make (related to DM module):
Could you provide some hints how to manage these problems? Thank you in advance for the help.
BTW, I noticed additional issues when using multi-processors (supposedly when creating documantation):