regression-test improvement in pre-release of PnetCDF 1.9.0

Sat Nov 4 12:51:46 CDT 2017

Hi Carl

The building/checking progress messages reported by automake and
libtool are in different form. To minimize the report on screen,
you can use commands below.
   make -s LIBTOOLFLAGS=--silent V=1
   make -s LIBTOOLFLAGS=--silent V=1 check

The MPI error messages you are seeing are generated from OpenMPI
and they are harmless as we discussed this in our previous emails.
So, it is safe to ignore them. The rest of messages contain no
errors generated from PnetCDF. If a fatal PnetCDF error occurs, the
make or make check command will stop and report the errors on
screen. For example,
   ...
   Making all in dispatchers
   make[3]: *** No rule to make target `file.c', needed by `file.lo'.  Stop.
   make[2]: *** [all] Error 2
   make[1]: *** [all-recursive] Error 1
   make: *** [all-recursive] Error 1

Let me know if you encounter this kind of errors.

Wei-keng

On Nov 4, 2017, at 2:03 AM, Carl Ponder wrote:

> On 11/01/2017 07:04 PM, Wei-keng Liao wrote:
>> We are pleased to announce the preview release of PnetCDF 1.9.0,
>> parallel-netcdf-1.9.0.pre1, is now available for download.
>> We are sincerely soliciting feedbacks from the PnetCDF community
>> for this test release. The tar balls can be downloaded from:
>> 
>> http://cucis.ece.northwestern.edu/projects/PnetCDF/download.html
>> 
>> 
>> Below is a brief list of noticeable changes.
>> 1. PnetCDF has migrated to use automake and libtool.
>> 2. Support shared libraries. (ABI version starts from 1:0:0).
>> 3. NC_MAX_VAR_DIMS, NC_MAX_DIMS, NC_MAX_ATTRS, NC_MAX_VARS have been
>>    lift to NC_MAX_INT for all classic formats
>> 4. Improved support for fill mode
>> 5. Support ARM architecture
>> 6. Update UTF8 encoding to use utf8proc v2.1.0 from the Julia Language project
>> 7. Build recipe for Unbuntu
>> 8. Address issues/bugs caused by MPI libraries (OpenMPI an MVAPICH)
>> 
>> Please note this is the first release of PnetCDF that officially supports
>> shared libraries. Problems may occur during the building and linking stages.
>> As always, please report problems encountered to this mailing list.
>> 
>> Detailed changes made since 1.8.1 are listed in
>> 
>> http://trac.mcs.anl.gov/projects/parallel-netcdf/browser/tags/v1-9-0-pre1/sneak_peek
> The pre-release fixes some error-messages that I'd been seeing with 1.8.1 and previous:
> ./nf_test -c    -d .
> ./nf_test       -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F77 ./nf_test for CDF-1                                ------ pass
> ./nf_test -c -2 -d .
> ./nf_test -2    -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F77 ./nf_test for CDF-2                                ------ pass
> ./nf_test -c -5 -d .
> ./nf_test -5    -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F77 ./nf_test for CDF-5                                ------ pass
> make[2]: Leaving directory `/cm/extra/apps/PNetCDF/1.8.1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf_test'
> 
> ./nf90_test -c    -d .
> ./nf90_test       -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F90 ./nf90_test for CDF-1                              ------ pass
> ./nf90_test -c -2 -d .
> ./nf90_test -2    -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F90 ./nf90_test for CDF-2                              ------ pass
> ./nf90_test -c -5 -d .
> ./nf90_test -5    -d .
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> MPI error (MPI_File_delete) : MPI_ERR_IO: input/output error
> *** TESTING F90 ./nf90_test for CDF-5                              ------ pass
> make[2]: Leaving directory `/cm/extra/apps/PNetCDF/1.8.1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf90_test'
> 
> This was using OpenMPI 1.10.7.
> I can't tell for sure that the equivalent set of tests is being performed, since the list is reported differently:
> make[3]: Entering directory `/cm/extra/apps/PNetCDF/1.9.0.pre1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf_test'
>   PPF77    nf_test.o
>   PPF77    nf_error.o
>   PPF77    test_read.o
>   PPF77    test_write.o
>   PPF77    util.o
>   CC       fortlib.o
> m4  -DPNETCDF -I../../m4 test_get.m4 >test_get.F
>   PPF77    test_get.o
> m4  -DPNETCDF -I../../m4 test_put.m4 >test_put.F
>   PPF77    test_put.o
> m4  -DPNETCDF -I../../m4 test_iget.m4 >test_iget.F
>   PPF77    test_iget.o
> m4  -DPNETCDF -I../../m4 test_iput.m4 >test_iput.F
>   PPF77    test_iput.o
>   F77LD    nf_test
> make[3]: Leaving directory `/cm/extra/apps/PNetCDF/1.9.0.pre1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf_test'
> 
> make[3]: Entering directory `/cm/extra/apps/PNetCDF/1.9.0.pre1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf90_test'
>   PPFC     nf90_test.o
>   PPFC     nf90_error.o
>   PPFC     test_read.o
>   PPFC     test_write.o
>   PPFC     util.o
>   CC       fortlib.o
> m4  -DPNETCDF -I../../m4 test_get.m4 >test_get.F90
>   PPFC     test_get.o
> m4  -DPNETCDF -I../../m4 test_put.m4 >test_put.F90
>   PPFC     test_put.o
> m4  -DPNETCDF -I../../m4 test_iget.m4 >test_iget.F90
>   PPFC     test_iget.o
> m4  -DPNETCDF -I../../m4 test_iput.m4 >test_iput.F90
>   PPFC     test_iput.o
>   FCLD     nf90_test
> make[3]: Leaving directory `/cm/extra/apps/PNetCDF/1.9.0.pre1/PGI-17.10_OpenMPI-1.10.7_CUDA-9.0.176/distro/test/nf90_test'
> 
> This email message is for the sole use of the intended recipient(s) and may contain confidential information.  Any unauthorized review, use, disclosure or distribution is prohibited.  If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.