1. Exascale FastForward
  2. ParallelIO

Source

Source viewDiff to previous
 
Information needs to go in proper documentation
​
State of dev - Nov. 5 2014
​
Key:
​
>> shell command
-- command output 
# or = textual comment
​
===== Yellowstone =====
===== Yellowstone_intel =====
​
>>module load python
>>python scripts/python/buildPio.py intel
​
--75% tests passed, 5 tests failed out of 20
​
--Total Test time (real) = 290.20 sec
​
--The following tests FAILED:
--          5 - pio_file_fail (Failed)
--          9 - ncdf_fail (Failed)
--Errors while running CTest
​
>>cd build_yellowstone_intel/examples/f03
​
>>bsub -Is -q caldera -n2 -PP93300070 -W0:20 "mpirun.lsf ./examplePio_f90"
>>ncdump examplePio_f90.nc | grep foo
        int foo(x) ;
 foo = 42, 42, 42, 42, 42, 42, 42, 42, 43, 43, 43, 43, 43, 43, 43, 43 ;
​
# or 
>>cd build_yellowstone_intel/examples/c
>>bsub -Is -q caldera -n8 -PP93300070 -W0:20 "mpirun.lsf ./examplePio_c"
--Job <115205> is submitted to queue <caldera>.
--<<Waiting for dispatch ...>>
--<<Starting on caldera15-ib>>
--Execute poe command line: poe  ./examplePio_c
--Running with 8 MPI processes and 8 PIO processes.
--Close file 65536
>>ncdump examplePio_c.nc | grep foo
--        int foo(x) ;
-- foo = 42, 42, 43, 43, 44, 44, 45, 45, 46, 46, 47, 47, 48, 48, 49, 49 ;
​
===== Goldbach (linux cluster) =====
===== Yellowstone_pgi =====
​
# no cxx build for examples since it seems like PGI's cxx compilers are broken on yellowstone?
​
>>module load python
>>python scripts/python/buildPio.py pgi
​
--21/21 Test #21: piotest ..........................***Failed    0.48 sec
​
--90% tests passed, 2 tests failed out of 21
​
--Total Test time (real) = 279.57 sec
​
--The following tests FAILED:
--          5 - pio_file_fail (Failed)
--         21 - piotest (Failed)
--Errors while running CTest
​
>>cd build_yellowstone_pgi/examples/f03
>>bsub -Is -q caldera -n8 -PP93300070 -W0:20 "mpirun.lsf ./examplePio_f90"
--Job <125873> is submitted to queue <caldera>.
--<<Waiting for dispatch ...>>
--<<Starting on caldera05-ib>>
--Execute poe command line: poe  ./examplePio_f90
--Close file 65536
>>ncdump examplePio_f90.nc
netcdf examplePio_f90 {
dimensions:
        x = 16 ;
variables:
        int foo(x) ;
data:
​
 foo = 42, 42, 43, 43, 44, 44, 45, 45, 46, 46, 47, 47, 48, 48, 49, 49 ;
}
​
# or
​
>>cd build_yellowstone_pgi/examples/c
>>ncdump examplePio_c.nc
--netcdf examplePio_c {
--dimensions:
--        x = 16 ;
--variables:
--        int foo(x) ;
--data:
​
-- foo = 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, _ ;
--}
​
#~# NOTE looks like a problem with PGI + C with 16 pes!  Compare with darwin_gnu
​
​
===== Yellowstone_gnu =====
​
​
​
​
===== Goldbach_nag (linux cluster) =====
​
# c examples won't build due to a lib_curl dependence for both intel and nag
​
>>python scripts/python/buildPio.py nag
--20/20 Test #20: pio_decomp_tests_4p_2iop_1agg ....   Passed    2.45 sec
​
--95% tests passed, 1 tests failed out of 20
​
--Total Test time (real) =  27.88 sec
​
--The following tests FAILED:
--          5 - pio_file_fail (Failed)
--Errors while running CTest
>>cd build_goldbach_nag/examples/f03
​
>>/usr/local/openmpi-1.6.5-gcc-g++-4.4.7-3-nag-5.3.1-907/bin/mpirun -np 4 ./examplePio_f90
Close file 65536
>>ncdump examplePio_f90.nc | grep foo
--        int foo(x) ;
-- foo = 42, 42, 42, 42, 43, 43, 43, 43, 44, 44, 44, 44, 45, 45, 45, 45 ;
​
​
​
​
===== Darwin =====
===== Goldbach_nag (linux cluster) =====
​
​
​
===== Darwin_gnu =====
​
# To run on a machine running OSX with gnu:
​
>>python scripts/python/buildPio.py gnu
​
# testing output ::
​
--The following tests FAILED:
--          5 - pio_file_fail (Failed)
--          6 - ncdf_simple_tests (Failed)