Skip to Content.
Sympa Menu

h5part - Re: [H5part] H5Hut fortran compiling

h5part AT lists.psi.ch

Subject: H5Part development and discussion

List archive

Re: [H5part] H5Hut fortran compiling


Chronological Thread  
  • From: "Gsell Achim (PSI)" <achim.gsell AT psi.ch>
  • To: "Koch, Benjamin" <benjamin.koch2 AT rwth-aachen.de>
  • Cc: "h5part AT lists.psi.ch" <h5part AT lists.psi.ch>
  • Subject: Re: [H5part] H5Hut fortran compiling
  • Date: Mon, 1 Mar 2021 15:39:56 +0000
  • Accept-language: en-US, de-CH
  • Authentication-results: localhost; iprev=pass (psi-seppmail1.ethz.ch) smtp.remote-ip=129.132.93.141; spf=pass smtp.mailfrom=psi.ch; dmarc=skipped

Hi Benjamin,

I just tried with GCC without any problems. With Intel I still have to test.


On 1 Mar 2021, at 11:56, Koch, Benjamin <benjamin.koch2 AT rwth-aachen.de> wrote:

I'm working on a linux cluster. 

Anything special on this cluster? Or just some standard Linux cluster @RWTH? 

Today, i tried it again, to have a clean environement.
Here is a list of how I installed H5hut, did I something wrong?

  1. git clone https://gitlab.psi.ch/H5Hut/src.git
  2. ./autogen.sh (in the src folder of course)
  3. ./configure --enable-fortran (Is this what you mean by how I call the configure script?)

Yes.

If you want to build the example, you can run configure with ‘--enable-examples’.

  1. make
  2. make test (Ouput: Nothing to do for test)
  3. Tried the tests in the test folder. The h5u_test gives an error for "read4" that the values are incorrect.

Which hdf5 version are you using? Please send the output of

h5cc -showconfig

and

h5pcc -showconfig


  1. make install

After that, I call al little script in the examples folder:
<Script.png>
In a similar way I was able to compile the C code, but maybe I'm missing something for Fortran?

In the script you are using the MPI wrapper of the Intel Fortran compiler but you compiled H5hut without MPI! To compile H5hut with MPI you have to run something like

configure --enable-parallel --enable-fortran ...

On some systems you have to set the environment variables CC and FC:
CC=mpiicc FC=mpiifort ./configure ...
make
...




Best

Achim


Best regards
Benjamin Koch

Of topic:
One could say, because the institute for Combustion Technology is known for the thermodynamics lecture. But there is also another Institute for technical thermodynamics.



Von: Gsell Achim (PSI) <achim.gsell AT psi.ch>
Gesendet: Freitag, 26. Februar 2021 13:29
An: Koch, Benjamin
Cc: h5part AT lists.psi.ch
Betreff: Re: [H5part] H5Hut fortran compiling
 
Hi Benjamin,

I just tried to reproduce on macOS 10.15 with gcc8 from Macports. For me everything compiled and linked fine.

On which system (Linux (distribution?), macOS, ...) are you compiling?

How do you call the configure script?

From where did you downloaded/cloned H5hut? The "official" URL is:

https://gitlab.psi.ch/H5hut/src


Best

Achim
--
Of topic: is the Institute for Combustion Technology the former Institut für Thermodynamik?


> On 25 Feb 2021, at 11:45, Koch, Benjamin <benjamin.koch2 AT rwth-aachen.de> wrote:
> 
> Hello team of H5Hut,
> 
> I'm a master student of the RWTH-Aachen and I work for the Institute for Combustion Technology. My supervisor and I are interested in data management via HDF5.
> We are working on the simulation of nano-particle combustion. Because the current method for writing out particle quantities could be faster, we discover your program on github.
> Our simulation program uses fortran code, so we try to include the fortran part of H5Hut.
> 
> I tried to run the examples and it worked for the C part of the code, but I can't get the fortran code working.
> 
> I think there is a problem with the linking of the library. The library seems to be linked, but then the error messages follow that some functions are not found. For example "pathto/src/build/lib/libH5hutF.so: undefined reference to `h5_flush_iteration'".
> 
> Before I get too deep into the solving this problem, I like to ask:
> The last updates of the fortran code seem to be some time ago, is it still up to date?
> 
> Best regards
> Benjamin Koch

Senior Scientist
Paul Scherrer Institut
Laboratory for Scientific Computing and Modeling
OHSA/D04
5232 Villigen PSI




Archive powered by MHonArc 2.6.19.

Top of Page