Garbled output on cluster

General discussion of all things GYRE-related (e.g., results, talks, ideas, tips)
Post Reply
warrick
Posts: 42
Joined: Wed Aug 28, 2013 2:47 am

Garbled output on cluster

Post by warrick » Wed May 30, 2018 10:39 am

Hi,

I'm bringing this to "General Discussion" because I don't think it really counts as a bug in GYRE. It's much more likely something strange about the computer setup. In short, the GYRE summary output is completely garbled. I was using my own fork of the code but I've checked that the same thing happens with GYRE 5.2 built using the MESA SDK 20180127. The computer part of our University computer cluster (BlueBEAR) and is running CentOS Linux release 7.5.1804 (Core) with a module system for packages. I've currently loaded the following modules, showing versions of some other packages:

Code: Select all

$ module list
Currently Loaded Modulefiles:
  1) dot                                                         12) numactl/2.0.11-GCC-5.4.0-2.26                               23) libreadline/6.3-foss-2016.10
  2) bb-variables/8                                              13) hwloc/1.11.5-GCC-5.4.0-2.26                                 24) Tcl/8.6.4-foss-2016.10
  3) slurm-moab-wrappers/2                                       14) OpenMPI/1.10.2-GCC-5.4.0-2.26                               25) SQLite/3.9.2-foss-2016.10
  4) slurm-helpers/3                                             15) gompi/2016.10                                               26) GMP/6.1.1-foss-2016.10
  5) openmpi-env/4                                               16) OpenBLAS/0.2.18-gompi-2016.10-LAPACK-3.6.1                  27) Tk/8.6.4-foss-2016.10
  6) handbuilt/1                                                 17) FFTW/3.3.4-gompi-2016.10                                    28) OpenSSL/1.0.2h-foss-2016.10
  7) bear-apps/2017a                                             18) ScaLAPACK/2.0.2-gompi-2016.10-OpenBLAS-0.2.18-LAPACK-3.6.1  29) Python/3.5.2-foss-2016.10-minimal
  8) bluebear/15                                                 19) foss/2016.10                                                30) numpy/1.13.0-foss-2016.10-Python-3.5.2
  9) GCCcore/5.4.0                                               20) bzip2/1.0.6-foss-2016.10                                    31) scipy/0.19.0-python-3.5.2
 10) binutils/2.26-GCCcore-5.4.0                                 21) zlib/1.2.8-foss-2016.10
 11) GCC/5.4.0-2.26                                              22) ncurses/6.0-foss-2016.10
The input and output files are attached.

For example, here's what I see using head:

Code: Select all

$ head gyre_summary.txt 
                                                                                                                                                                                                                                                                
                        1                        2                        3
                   M_star                   R_star                   L_star
  0.218  0.244  0.241
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
                              0.864407094  0.495
                              0.110337475  0.101
                              0.1343E+003  0.232
                              0.156  0.000  0.607
Worse still, if I use less, I just get nonsense:

Code: Select all

                        1                        2                        3
                   M_star                   R_star                   L_star
 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@1^@^@^@8 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@4^@^@^@4 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@4^@^@^@1
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@8^@^@^@6^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@1^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@3^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@5^@
...
Running on my personal desktop (running Fedora 28), I get, for example,

Code: Select all

$ head gyre_summary.txt
                                                                                                                                                                                                                                                                
                        1                        2                        3
                   M_star                   R_star                   L_star
  0.2186687775470000E+034  0.2447592756970000E+012  0.2411088716180000E+035
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
                        0                        3                        3                        0  0.8641133813210968E+002  0.0000000000000000E+000  0.4959470410287945E-003
                        0                        4                        4                        0  0.1107188819005652E+003  0.0000000000000000E+000  0.1017498891469212E-003
                        0                        5                        5                        0  0.1343250413866053E+003  0.0000000000000000E+000  0.2321497288132400E-004
                        0                        6                        6                        0  0.1563790964964384E+003  0.0000000000000000E+000  0.6073443783042598E-005
Any ideas? I can provide details about the cluster setup but I'm not sure what's relevant.

Cheers,
Warrick
Attachments
gyre_summary.txt
(3.91 KiB) Downloaded 57 times
gyre.in
(946 Bytes) Downloaded 47 times
final.profile.GYRE
(780.01 KiB) Downloaded 55 times

User avatar
rhtownsend
Site Admin
Posts: 284
Joined: Sun Mar 31, 2013 4:22 pm

Re: Garbled output on cluster

Post by rhtownsend » Wed May 30, 2018 10:44 am

warrick wrote:
Wed May 30, 2018 10:39 am
Hi,

I'm bringing this to "General Discussion" because I don't think it really counts as a bug in GYRE. It's much more likely something strange about the computer setup. In short, the GYRE summary output is completely garbled. I was using my own fork of the code but I've checked that the same thing happens with GYRE 5.2 built using the MESA SDK 20180127. The computer part of our University computer cluster (BlueBEAR) and is running CentOS Linux release 7.5.1804 (Core) with a module system for packages. I've currently loaded the following modules, showing versions of some other packages:

Code: Select all

$ module list
Currently Loaded Modulefiles:
  1) dot                                                         12) numactl/2.0.11-GCC-5.4.0-2.26                               23) libreadline/6.3-foss-2016.10
  2) bb-variables/8                                              13) hwloc/1.11.5-GCC-5.4.0-2.26                                 24) Tcl/8.6.4-foss-2016.10
  3) slurm-moab-wrappers/2                                       14) OpenMPI/1.10.2-GCC-5.4.0-2.26                               25) SQLite/3.9.2-foss-2016.10
  4) slurm-helpers/3                                             15) gompi/2016.10                                               26) GMP/6.1.1-foss-2016.10
  5) openmpi-env/4                                               16) OpenBLAS/0.2.18-gompi-2016.10-LAPACK-3.6.1                  27) Tk/8.6.4-foss-2016.10
  6) handbuilt/1                                                 17) FFTW/3.3.4-gompi-2016.10                                    28) OpenSSL/1.0.2h-foss-2016.10
  7) bear-apps/2017a                                             18) ScaLAPACK/2.0.2-gompi-2016.10-OpenBLAS-0.2.18-LAPACK-3.6.1  29) Python/3.5.2-foss-2016.10-minimal
  8) bluebear/15                                                 19) foss/2016.10                                                30) numpy/1.13.0-foss-2016.10-Python-3.5.2
  9) GCCcore/5.4.0                                               20) bzip2/1.0.6-foss-2016.10                                    31) scipy/0.19.0-python-3.5.2
 10) binutils/2.26-GCCcore-5.4.0                                 21) zlib/1.2.8-foss-2016.10
 11) GCC/5.4.0-2.26                                              22) ncurses/6.0-foss-2016.10
The input and output files are attached.

For example, here's what I see using head:

Code: Select all

$ head gyre_summary.txt 
                                                                                                                                                                                                                                                                
                        1                        2                        3
                   M_star                   R_star                   L_star
  0.218  0.244  0.241
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
                              0.864407094  0.495
                              0.110337475  0.101
                              0.1343E+003  0.232
                              0.156  0.000  0.607
Worse still, if I use less, I just get nonsense:

Code: Select all

                        1                        2                        3
                   M_star                   R_star                   L_star
 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@1^@^@^@8 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@4^@^@^@4 ^@^@^@ ^@^@^@0^@^@^@.^@^@^@2^@^@^@4^@^@^@1
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@8^@^@^@6^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@1^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@3^@
 ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@ ^@^@^@  ^@^@^@ ^@^@^@0^@^@^@.^@^@^@1^@^@^@5^@
...
Running on my personal desktop (running Fedora 28), I get, for example,

Code: Select all

$ head gyre_summary.txt
                                                                                                                                                                                                                                                                
                        1                        2                        3
                   M_star                   R_star                   L_star
  0.2186687775470000E+034  0.2447592756970000E+012  0.2411088716180000E+035
                        1                        2                        3                        4                        5                        6                        7
                        l                     n_pg                      n_p                      n_g                 Re(freq)                 Im(freq)                   E_norm
                        0                        3                        3                        0  0.8641133813210968E+002  0.0000000000000000E+000  0.4959470410287945E-003
                        0                        4                        4                        0  0.1107188819005652E+003  0.0000000000000000E+000  0.1017498891469212E-003
                        0                        5                        5                        0  0.1343250413866053E+003  0.0000000000000000E+000  0.2321497288132400E-004
                        0                        6                        6                        0  0.1563790964964384E+003  0.0000000000000000E+000  0.6073443783042598E-005
Any ideas? I can provide details about the cluster setup but I'm not sure what's relevant.

Cheers,
Warrick
Hi Warrick --

What happens if you copy the summary.txt file from the cluster to your personal desktop, and then use head/less there? Is the file intrinsically wrong, or is it just displaying wrong? The latter would suggest that the cluster's terminal settings may have an issue.

cheers,

Rich

warrick
Posts: 42
Joined: Wed Aug 28, 2013 2:47 am

Re: Garbled output on cluster

Post by warrick » Wed May 30, 2018 10:54 am

No change. If I copy the file to my desktop and view it, I see the same thing.

FWIW, the files do have the same size (4005 bytes).

User avatar
rhtownsend
Site Admin
Posts: 284
Joined: Sun Mar 31, 2013 4:22 pm

Re: Garbled output on cluster

Post by rhtownsend » Wed May 30, 2018 10:57 am

warrick wrote:
Wed May 30, 2018 10:54 am
No change. If I copy the file to my desktop and view it, I see the same thing.

FWIW, the files do have the same size (4005 bytes).
Could you post both files to the forum, so I can examine in detail?

Also, could you cut and paste (and post) the output from 'ldd gyre', on the cluster.

cheers,

Rich

warrick
Posts: 42
Joined: Wed Aug 28, 2013 2:47 am

Re: Garbled output on cluster

Post by warrick » Wed May 30, 2018 11:06 am

Hi Rich,

Sure! The two files are attached as "good" (on my desktop, which is what I expect from GYRE) and "bad".

And here's the terminal output:

Code: Select all

$ ldd $GYRE_DIR/bin/gyre
	linux-vdso.so.1 =>  (0x00007ffcba85f000)
	libhdf5_fortran.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5_fortran.so.10 (0x00007f9c2cce4000)
	libhdf5.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5.so.10 (0x00007f9c2c7c6000)
	libz.so.1 => /rds/bear-apps/2017a/EL7-sandybridge/software/zlib/1.2.8-foss-2016.10/lib/libz.so.1 (0x00007f9c2c5b0000)
	liblapack95.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack95.so (0x00007f9c2c1e1000)
	liblapack.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack.so (0x00007f9c2b872000)
	libxblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libxblas.so (0x00007f9c2b1da000)
	libblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libblas.so (0x00007f9c2af4b000)
	libgfortran.so.3 => /rds/bear-apps/2017a/EL7-sandybridge/software/GCCcore/5.4.0/lib64/libgfortran.so.3 (0x00007f9c2ac2b000)
	libm.so.6 => /usr/lib64/libm.so.6 (0x00007f9c2a915000)
	libgomp.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgomp.so.1 (0x00007f9c2a6e5000)
	libgcc_s.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgcc_s.so.1 (0x00007f9c2a4ce000)
	libquadmath.so.0 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libquadmath.so.0 (0x00007f9c2a28d000)
	libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007f9c2a071000)
	libc.so.6 => /usr/lib64/libc.so.6 (0x00007f9c29ca4000)
	librt.so.1 => /usr/lib64/librt.so.1 (0x00007f9c29a9b000)
	libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007f9c29897000)
	libgfortran.so.4 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/../lib64/libgfortran.so.4 (0x00007f9c2954b000)
	libtmg.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libtmg.so (0x00007f9c292e5000)
	/lib64/ld-linux-x86-64.so.2 (0x00007f9c2cf33000)
Attachments
gyre_summary_good.txt
(3.91 KiB) Downloaded 37 times

warrick
Posts: 42
Joined: Wed Aug 28, 2013 2:47 am

Re: Garbled output on cluster

Post by warrick » Wed May 30, 2018 11:12 am

Okay, the problem is almost certainly in the module system somewhere. I just reset everything to system defaults and loaded just the SDK, which now gives the expected output file. Here's that ldd output when it works correctly. No time now but I'll cross check to see if the troublesome package turns up.

Code: Select all

$ ldd $GYRE_DIR/bin/gyre
	linux-vdso.so.1 =>  (0x00007ffd75b56000)
	libhdf5_fortran.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5_fortran.so.10 (0x00007fb2de274000)
	libhdf5.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5.so.10 (0x00007fb2ddd55000)
	libz.so.1 => /usr/lib64/libz.so.1 (0x00007fb2ddb2c000)
	liblapack95.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack95.so (0x00007fb2dd75d000)
	liblapack.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack.so (0x00007fb2dcded000)
	libxblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libxblas.so (0x00007fb2dc756000)
	libblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libblas.so (0x00007fb2dc4c7000)
	libgfortran.so.4 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgfortran.so.4 (0x00007fb2dc17a000)
	libm.so.6 => /usr/lib64/libm.so.6 (0x00007fb2dbe78000)
	libgomp.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgomp.so.1 (0x00007fb2dbc48000)
	libgcc_s.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgcc_s.so.1 (0x00007fb2dba30000)
	libquadmath.so.0 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libquadmath.so.0 (0x00007fb2db7f0000)
	libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007fb2db5d4000)
	libc.so.6 => /usr/lib64/libc.so.6 (0x00007fb2db206000)
	librt.so.1 => /usr/lib64/librt.so.1 (0x00007fb2daffe000)
	libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007fb2dadfa000)
	libtmg.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libtmg.so (0x00007fb2dab94000)
	/lib64/ld-linux-x86-64.so.2 (0x00007fb2de4c2000)
Thanks for the speedy replies, which prompted me to resetting the system (which I should've done in the first place). I'll play around to make sure I can consistently reproduce this before I bother you again!

User avatar
rhtownsend
Site Admin
Posts: 284
Joined: Sun Mar 31, 2013 4:22 pm

Re: Garbled output on cluster

Post by rhtownsend » Wed May 30, 2018 11:27 am

warrick wrote:
Wed May 30, 2018 11:12 am
Okay, the problem is almost certainly in the module system somewhere. I just reset everything to system defaults and loaded just the SDK, which now gives the expected output file. Here's that ldd output when it works correctly. No time now but I'll cross check to see if the troublesome package turns up.

Code: Select all

$ ldd $GYRE_DIR/bin/gyre
	linux-vdso.so.1 =>  (0x00007ffd75b56000)
	libhdf5_fortran.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5_fortran.so.10 (0x00007fb2de274000)
	libhdf5.so.10 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libhdf5.so.10 (0x00007fb2ddd55000)
	libz.so.1 => /usr/lib64/libz.so.1 (0x00007fb2ddb2c000)
	liblapack95.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack95.so (0x00007fb2dd75d000)
	liblapack.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/liblapack.so (0x00007fb2dcded000)
	libxblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libxblas.so (0x00007fb2dc756000)
	libblas.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libblas.so (0x00007fb2dc4c7000)
	libgfortran.so.4 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgfortran.so.4 (0x00007fb2dc17a000)
	libm.so.6 => /usr/lib64/libm.so.6 (0x00007fb2dbe78000)
	libgomp.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgomp.so.1 (0x00007fb2dbc48000)
	libgcc_s.so.1 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libgcc_s.so.1 (0x00007fb2dba30000)
	libquadmath.so.0 => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib64/libquadmath.so.0 (0x00007fb2db7f0000)
	libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007fb2db5d4000)
	libc.so.6 => /usr/lib64/libc.so.6 (0x00007fb2db206000)
	librt.so.1 => /usr/lib64/librt.so.1 (0x00007fb2daffe000)
	libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007fb2dadfa000)
	libtmg.so => /rds/projects/2017/ballwh-stellar-modelling/mesa/sdk/20180127/lib/libtmg.so (0x00007fb2dab94000)
	/lib64/ld-linux-x86-64.so.2 (0x00007fb2de4c2000)
Thanks for the speedy replies, which prompted me to resetting the system (which I should've done in the first place). I'll play around to make sure I can consistently reproduce this before I bother you again!
Looking at the previous ldd output, I see that libgfortran.so.3 is getting pulled in, from outside the SDK. Given that libgfortran handles the Fortran I/O, I'm guessing that it is the culprit.

cheers,

Rich

Post Reply