I'm sorry Mengda, but we don't support Windows anymore. You can find out more on our FAQ, but in short, you'll need to try another implementation such as Microsoft MPI or Intel MPI. Thanks, Wesley

On Sep 20, 2014, at 9:50 PM, Mengda Wu <[email protected]> wrote:

Hello all,

I am trying to compile mpich-3.1.2 under Cygwin 64bit. I am using the mingw64- x86_64-gcc-g++ and mingw64-x86_64-gcc-gfortran compilers, i.e. the mingw compilers shipped in the latest Cygwin. I encountered an internal error "Failed to convert MPI_INTEGER value to hex!" when configuring.

======================================================== checking whether real*4 is supported... yes checking whether real*8 is supported... yes checking whether real*16 is supported... yes : integer expression expected: 4 : integer expression expected: 4 : integer expression expected: 4 : integer expression expected: 8 : integer expression expected: 8 : integer expression expected: 8 checking for C type matching Fortran real... unavailable checking for C type matching Fortran double... unavailable expr: syntax error expr: syntax error expr: non-integer argument ./configure: line 38715: test: : integer expression expected ./configure: line 38720: test: -gt: unary operator expected ./configure: line 38734: test: -lt: unary operator expected ./configure: line 38715: test: : integer expression expected ./configure: line 38720: test: -gt: unary operator expected ./configure: line 38734: test: -lt: unary operator expected ./configure: line 38715: test: : integer expression expected ./configure: line 38720: test: -gt: unary operator expected ./configure: line 38734: test: -lt: unary operator expected expr: non-integer argument ./configure: line 38783: test: -ge: unary operator expected expr: syntax error expr: syntax error expr: syntax error expr: syntax error configure: error: Internal Error: Failed to convert MPI_INTEGER value to hex! O riginal value was 0x4c001b =============================================================

In config.log, it shows ============================================================= configure:38248: checking for size of Fortran type double precision configure:38289: gcc -c -O2 -I/cygdrive/c/Library/mpich2/mpich-3.1.2/src/mpl/include -I/cygdrive/c/Library/mpich2/mpich-3.1.2/src/mpl/include -I/cygdrive/c/Library/mpich2/mpich-3.1.2/src/openpa/src -I/cygdrive/c/Library/mpich2/mpich-3.1.2/src/openpa/src -D_REENTRANT -I/cygdrive/c/Library/mpich2/mpich-3.1.2/src/mpi/romio/include conftest.c >&5 configure:38289: $? = 0 configure:38293: mv conftest.o pac_conftest.o configure:38296: $? = 0 configure:38323: gfortran -o conftest.exe -O2 conftest.f pac_conftest.o >&5 configure:38323: $? = 0 configure:38323: ./conftest.exe configure:38323: $? = 0 configure:38361: result: 8

configure:38384: checking whether integer*1 is supported configure:38395: gfortran -c -O2 conftest.f >&5 configure:38395: $? = 0 configure:38402: result: yes configure:38404: checking whether integer*2 is supported configure:38415: gfortran -c -O2 conftest.f >&5 configure:38415: $? = 0 configure:38422: result: yes configure:38424: checking whether integer*4 is supported configure:38435: gfortran -c -O2 conftest.f >&5 configure:38435: $? = 0 configure:38442: result: yes configure:38444: checking whether integer*8 is supported configure:38455: gfortran -c -O2 conftest.f >&5 configure:38455: $? = 0 configure:38462: result: yes configure:38464: checking whether integer*16 is supported configure:38475: gfortran -c -O2 conftest.f >&5 configure:38475: $? = 0 configure:38482: result: yes configure:38484: checking whether real*4 is supported configure:38495: gfortran -c -O2 conftest.f >&5 configure:38495: $? = 0 configure:38502: result: yes configure:38504: checking whether real*8 is supported configure:38515: gfortran -c -O2 conftest.f >&5 configure:38515: $? = 0 configure:38522: result: yes configure:38524: checking whether real*16 is supported configure:38535: gfortran -c -O2 conftest.f >&5 configure:38535: $? = 0 configure:38542: result: yes configure:38641: checking for C type matching Fortran real configure:38659: result: unavailable configure:38664: checking for C type matching Fortran double configure:38681: result: unavailable configure:38842: error: Internal Error: Failed to convert MPI_INTEGER value to hex! Original value was 0x4c001b =============================================================

Can you help?

Thanks, Mengda

_______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss