ICE caused by Blitz++ example program

Martin Reinecke martin@MPA-Garching.MPG.DE
Thu Jul 1 00:56:00 GMT 1999


Hi,

When compiling the an example from the Blitz++ library (snapshot 1990629)
with gcc-2.95 19990615 on AIX 4.2, I get the following error message:


~/download/blitz-19990629/examples>g++ -v --save-temps  -ftemplate-depth-30 -O2 -funroll-loops -fstrict-aliasing -fno-gcse -g -DBZ_DEBUG -I./.. -c ./cfd.cpp
Reading specs from /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/specs
gcc version gcc-2.95 19990615 (prerelease)
 /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/cpp -lang-c++ -v -I./.. -D__GNUC__=2 -D__GNUG__=2 -D__GNUC_MINOR__=95 -D__cplusplus
-D_IBMR2 -D_POWER -D_AIX -D_AIX32 -D_AIX41 -D_LONG_LONG -D_IBMR2 -D_POWER -D_AIX -D_AIX32 -D_AIX41 -D_LONG_LONG -Asystem(unix) -Asystem(aix) -D__EXCEPTIONS
-D__CHAR_UNSIGNED__ -D__OPTIMIZE__ -g -D_ARCH_COM -DBZ_DEBUG ./cfd.cpp cfd.ii
GNU CPP version gcc-2.95 19990615 (prerelease)
#include "..." search starts here:
#include <...> search starts here:
 ..
 /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/../../../../include/g++-2
 /usr/local/include
 /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/../../../../powerpc-ibm-aix4.2.1.0/include
 /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/include
 /usr/include
End of search list.
The following default directories have been omitted from the search path:
End of omitted list.
 /afs/mpa/data/agroebl/martin/egcsbeta/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/gcc-2.95/cc1plus cfd.ii -quiet -dumpbase cfd.cc -g -O2 -version -ftemplate-depth-30 -funroll-loops
-fstrict-aliasing -fno-gcse -o cfd.s
GNU C++ version gcc-2.95 19990615 (prerelease) (powerpc-ibm-aix4.2.1.0) compiled by GNU C version gcc-2.95 19990615 (prerelease).
In file included from ../blitz/numinquire.h:52,
                 from ../blitz/array/expr.h:55,
                 from ../blitz/array.h:2460,
                 from ./cfd.cpp:36:
../blitz/limits-hack.h:291: warning: large integer implicitly truncated to unsigned type
../blitz/limits-hack.h:291: warning: large integer implicitly truncated to unsigned type
../blitz/array/cgsolve.h: In function `int blitz::conjugateGradientSolver<P_solver_update, double, 3, PressureBCs>(P_solver_update, blitz::Array<double,3> &,
blitz::Array<double,3> &, double, const PressureBCs &)':
./cfd.cpp:320:   instantiated from here
../blitz/array/cgsolve.h:121: Internal compiler error.
../blitz/array/cgsolve.h:121: Please submit a full bug report to `egcs-bugs@egcs.cygnus.com'.
../blitz/array/cgsolve.h:121: See <URL: http://egcs.cygnus.com/faq.html#bugreport > for details.


The preprocessed source file is available at 
http://www.mpa-garching.mpg.de/~martin/cfd.ii.gz

If you need more information, please ask!


Martin Reinecke


More information about the Gcc-bugs mailing list