Arch Linux User Repository

@blixawillbargeld @mauritiusdadd With the current PKGBUILD only "symmetric stiffness/mass contributions" is calculated using multithreading in static analysis with spooles. The AUR Package spooles is compiled for MT. Adding -DUSE_MT to the CFLAGS in the Makefile of calculix enables multithreading for "Factoring the system of equations" with SPOOLES solver referring to http://www.dhondt.de/ccx_2.17.README.INSTALL Successfully tested with changed PKGBUILD and the following notice:

at execution time the environment variable OMP_NUM_THREADS must have been set to the number of CPUs you want to use. In Linux this can be done by ``export OMP_NUM_THREADS=n'' on the command line, where n is the number of CPUs. Default is 1. Alternatively, you can set the number of CPUs using the environment variable CCX_NPROC_EQUATION_SOLVER. If both are set, the latter takes precedence. Source: http://web.mit.edu/calculix_v2.7/CalculiX/ccx_2.7/doc/ccx/node3.html

The ccx manual has much more information about this topic: http://www.dhondt.de/ccx_2.17.pdf Page 12

Hey @mauritiusdadd , upgrading GCC seems to have fixed the problem. I'm running Manjaro and just found out that GCC is apparently one of the few packages that's treated differently in Manjaro and Arch. Had to move my system to testing repos in order to install GCC 10. Thanks for the help!!

Hi @steveo_155, is your system fully updated? In particular, do you have gcc 10?

I am getting the following build error:

gfortran -Wall -O2 -fallow-argument-mismatch -c absolute_relative.f gfortran: error: unrecognized command line option ‘-fallow-argument-mismatch’; did you mean ‘-Wno-argument-mismatch’? make: *** [Makefile:11: absolute_relative.o] Error 1 ==> ERROR: A failure occurred in build(). Aborting... Error making: calculix

@dkcampbell thank you very much! I'have been quite busy in the last couple of days. I'll merge your patch by the end of tomorrow, thanks again.

I have problem then compiling ccx (MKL)

sgetv0.f:404:32:

  367 |           call svout (logfil, 1, rnorm0, ndigit,
      |                                 2
......
  404 |          call svout (logfil, n, resid, ndigit,
      |                                1
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
make[1]: *** [Makefile:50: sgetv0.o] Błąd 1
make[1]: Opuszczenie katalogu '/home/ProjektK66/ccx/src/x64/build/ARPACK/SRC'
ranlib /home/ProjektK66/ccx/src/x64/build/ARPACK/libarpack_MINGW64.a
...
gfortran -Wall -O2 -fopenmp -c allocation.f
allocation.f:1634:72:

 1634 |      &            "*NODE or *NODE PRINT or *NODE FILE or *NODE OUTPUT%")
      |                                                                        1
Error: Missing actual argument for argument '_formal_15' at (1)

Im add -fallow-argument-mismatch to Makefile ( https://paste.sr.ht/~dkc/2f2ab3b277dfff53a06cf9162eea067dea0b7988 )

Hi @mauritiusdadd, the build issues @crataegus and myself are experiencing is caused by changes to the default options used by gcc, and gfortran with the release of gcc 10. For gfortran, a warning, by default is now an error, and for gcc, an ignored warning, became an error. Here's a link to a patch to fix the issue: https://paste.sr.ht/%7Edkc/2f2ab3b277dfff53a06cf9162eea067dea0b7988

I applied it on top of your patch. Feel free to merge this into your existing patch.

Thanks for your work maintaining this package!

@crataegus and @dkcampbell thanks for reporting this, last time I have checked the package it was building correctly in a clean chroot. I will investigate on this asap.