- Dec 01, 2022
-
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
Both 'head -n 1' and 'sed 1q' exit after printing the first line, which closes the pipe and kills the compiler's process leaving it no chance to clean any temporary files it might have created (e.g. Intel compiler creates 'a.out' when it is called with additional flags like -I, -L, etc., which is the case when $CC/$FC is an MPI wrapper.
-
Sergey Kosukhin authored
And make sure that the result is the same as for Fortran (FC).
-
Sergey Kosukhin authored
The conditional indicates that FC is available and all required compiler flags are known.
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
Instead of tinkering its analogue.
-
-
-
-
-
-
-
-
-
-
* Given how many lines documenting the confusing pattern and how many compiler warnings are eliminated by this change...
-
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
The ONLY clause works only with MPI versions that have all MPI subroutines declared in the module. According to Thomas Jahns, this is not always the case (see comments in #4 for more details).
-
Sergey Kosukhin authored
This way, we avoid target-specific flags '-DUSE_MPI', which result into target-specific rules that do not use Libtool for the compilation.
-
Sergey Kosukhin authored
Remove declaration of the 'BUILD_C_INCLUDE' and 'BUILD_FC_INCLUDE' precious variables from the configure script since they are not used.
-
Sergey Kosukhin authored
The shortcut for the ISO C99 test, which simplifies the implementation, is copied from PPM.
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-
Sergey Kosukhin authored
-