Skip to content
Snippets Groups Projects
Select Git revision
  • b5cddf4344cfa18f08afbd0ffab7cd127ef5ee91
  • master default protected
  • compilation2022apr
  • ISEI_3_5_1
  • VERSION_3_9-alba
  • VERSION_3_9-Indus2
  • Jianfeng
  • VERSION-3_10
  • VERSION-3_9_1
  • VERSION-3_9_alba
  • VERSION-3_9_Indus2
  • VERSION-3_9
  • VERSION-3_8
  • VERSION-3_7
  • ISEI_3_5_1-PATCH_2
  • ISEI_3_5_1-PATCH_1
  • PROD_3_5_1
  • VERSION_3_6prerelease2
  • VERSION_3_6prerelease
  • VERSION-3_5
  • tracy
21 results

notes

Blame
  • Code owners
    Assign users and groups as approvers for specific file changes. Learn more.
    notes 2.81 KiB
    # the following are some notes about the compilation of tracy on sumo-metis and
    # the CCRT for SOLEIL
    # orblancog
    # 2023aug28
    
    # First of all, let me apologize as this is not a script, but some notes you
    # will need to read and understand in order to build tracy.
    # Tracy depends on external non-standard libraries, and because of it and the
    # lack of a git repository for the external libraries, we will need to get a
    # copy from Laurent Nadolsky.
    
    
    # There two main parts :
      - numerical recipes (ask Laurent)
      - tracy3 (on gitlab)
    
    If you have numerical recipes and tracy inside the same folder, you could just
    copy 'buildtracy.sh' to that folder and launch it. There is one for sumo-metis
    and one for the CCRT.
    The first time 'buildtracy.sh' is launch it might stop due to a problem with
    make distclean inside make_for_tracy3.h. My advice is to comment that line
    in the first compilation, which would reset the project, and decomment that line
    for the next compilations.
    The goal with 'buildtracy.sh' is to produce a folder called 'build' with the
    binaries for a new module.
    
    Most of the time the above is enough. However, here below there is more detail.
    
    I have added a script 'set-vars-sumo-metis.sh' that creates few environment
    variables that are common to both numerical recipes and tracy3
    during compilation.
    My advise is to source it.
    
    --- NUMERICAL RECIPES
    The code has at least two versions. We will use here the c-ansi.
    On sumo metis:
       The compiler will be gcc because icc is incompatible with matlab.
       See jira ticket https://jira.synchrotron-soleil.fr/jira/browse/CTRLDESK-21799
       We should load tools/gcc/9.2.0, (it is a module visible from sumo or sumo-metis).
       In numerical_recipes/c-ansi/ there is a Makefile.
       I leave here a copy of that Makefile "Makefile_numerical_recipes_c-ansi".
       It seems to me that the Makefile in (numerical_recipes_c-ansi)/recipes was also
       modified sometime ago, I will leave a copy of it "Makefile_numerical_recipes_c-ansi_recipes".
    On the CCRT :
       The compiler is gnu/11.1.0 because we want to keep compatibility with elegant
       I had to modify the Makefile in the CCRT to make it work with the gcc
       version 11. I leave here the copies "CCRT_Makefile_numerical_recipes_c-ansi", and
       "CCRT_Makefile_numerical_recipes_c-ansi_recipes".
       I also needed to modify anneal.c to to use new libraries instead of deprecated no
       available in gcc ([TGCC #135313] /usr/include/bits/huge_val.h);
       I leave here a copy "CCRT_anneal.c".
    
    --- tracy3
    On sumo metis:
       Check the branch compilation2022apr of this project
       https://gitlab.synchrotron-soleil.fr/PA/optics/tracy3/-/commits/compilation2022apr
       It contains the most up-to-date modifications of the code
    On the CCRT:
       I leave a static copy of the code at
       /nfs/ruche/sources/physmach/blanco-garcia/backup_CCRT/work/public/progs/tracy3soleil/compileTracy