[Users] Issues compiling ETK on Blue Waters

[Federico Cipolletta]

I am experiencing some issues compiling the ETK on BW.

The command I use is the standard command for simfactory, where I specify
the machine as follows:

./simfactory/bin/sim build ETK_standard --mdbkey make 'make -j16'
--machine=bluewaters --thornlist=../einsteintoolkit.th | cat >&
compilation_trial.out

The error on which I am actually stuck is the following:

COMPILING arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:112:48:
error: macro "CCTK_LOOP3STROFF_NORMAL" requires 24 arguments, but only 23
given
              dstipadext, dstjpadext, dstkpadext) {
                                                ^
In file included from
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_core.h:157:0,
                 from
/mnt/a/u/sciteam/cipollet/Cactus/configs/ETK_standard/bindings/include/CarpetLib/cctk.h:7,
                 from
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:1:
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:
In function 'void CarpetLib::interpolate_3d_2tl(const T*, CCTK_REAL8, const
T*, CCTK_REAL8, const ivect3&, const ivect3&, T*, CCTK_REAL8, const
ivect3&, const ivect3&, const ibbox3&, const ibbox3&, const ibbox3&, const
ibbox3&, void*)':
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2460:3: error:
'CCTK_LOOP3STROFF_NORMAL' was not declared in this scope
   CCTK_LOOP3STROFF_NORMAL(name, \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2444:3: note: in
expansion of macro 'CCTK_LOOP3STROFF'
   CCTK_LOOP3STROFF(name, \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2428:3: note: in
expansion of macro 'CCTK_LOOP3STR'
   CCTK_LOOP3STR(name, \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:111:3:
note: in expansion of macro 'CCTK_LOOP3'
   CCTK_LOOP3(interpolate_3d_2tl, i, j, k, 0, 0, 0, regiext, regjext,
regkext,
   ^
In file included from
/mnt/a/u/sciteam/cipollet/Cactus/configs/ETK_standard/bindings/include/loopcontrol.h:4:0,
                 from
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:4:
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:
At global scope:
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/LoopControl/src/loopcontrol.h:263:3:
error: expected declaration before '}' token
   }
     \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/LoopControl/src/loopcontrol.h:304:42:
note: in expansion of macro 'LC_ENDLOOP3STROFF_NORMAL'
 #define CCTK_ENDLOOP3STROFF_NORMAL(name) LC_ENDLOOP3STROFF_NORMAL(name)
                                          ^
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2470:3: note: in
expansion of macro 'CCTK_ENDLOOP3STROFF_NORMAL'
   CCTK_ENDLOOP3STROFF_NORMAL(name) \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2452:3: note: in
expansion of macro 'CCTK_ENDLOOP3STROFF'
   CCTK_ENDLOOP3STROFF(name) \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/src/include/cctk_Loop.h:2436:3: note: in
expansion of macro 'CCTK_ENDLOOP3STR'
   CCTK_ENDLOOP3STR(name) \
   ^
/mnt/a/u/sciteam/cipollet/Cactus/arrangements/Carpet/CarpetLib/src/interpolate_3d_2tl.cc:116:3:
note: in expansion of macro 'CCTK_ENDLOOP3'
   CCTK_ENDLOOP3(interpolate_3d_2tl);
   ^
make[3]: *** [interpolate_3d_2tl.cc.o] Error 1
make[2]: *** [make.checked] Error 2
make[1]: ***
[/mnt/a/u/sciteam/cipollet/Cactus/configs/ETK_standard/lib/libthorn_CarpetLib.a]
Error 2
make: *** [ETK_standard] Error 2

As you can see, the CarpetLib is not compiled, due to an error on a call to
a macro of Cactus. As the CarpetLib is a pretty standard one, I assume that
there should be some bugs in the machine configuration file in simfactory,
probably when specifying the flags for the compilers. I also tried the
other configuration for BW machine (i.e. the "-cray" and "-pgi" versions)
when trying to compile another config of the ETK on BW, but neither those
where succesful.

Can anyone give me some hint?

Thank you very much,
Federico Cipolletta.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.einsteintoolkit.org/pipermail/users/attachments/20190814/79f7dce4/attachment.html