sddssort compiled from sources does not sort

[blanco-garcia]

Dear all,

I'm compiling the SDDS sources for a CENTOS 7 and CENTOS 8 x86_64 machine, but, although binaries are produced I see a strange behaviour.
Here below is the output I get from sddssort. Is there any particular parameter that should be passed, appart from make MPI=1 ?
When doing ldd I see nothing missing.

$ldd $(which sddssort)
linux-vdso.so.1 => (0x00007ffdd4387000)
intel64_lin/libmkl_def.so (0x00007fce1df35000)
libmkl_avx.so (0x00007fce1b89c000)
libmkl_avx2.so (0x00007fce18815000)
libmkl_core.so (0x00007fce16d22000)
libmkl_intel_lp64.so (0x00007fce162e2000)
libmkl_intel_ilp64.so (0x00007fce15958000)
intel64_lin/libmkl_intel_thread.so (0x00007fce13ec6000)
intel64_lin/libiomp5.so (0x00007fce13b24000)
libz.so.1 => /lib64/libz.so.1 (0x00007fce1390e000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fce136f2000)
libreadline.so.6 => /lib64/libreadline.so.6 (0x00007fce134ac000)
librt.so.1 => /lib64/librt.so.1 (0x00007fce132a4000)
libdl.so.2 => /lib64/libdl.so.2 (0x00007fce130a0000)
libstdc++.so.6 => /lib64/libstdc++.so.6 (0x00007fce12d99000)
libm.so.6 => /lib64/libm.so.6 (0x00007fce12a97000)
libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007fce12881000)
libc.so.6 => /lib64/libc.so.6 (0x00007fce124b3000)
/lib64/ld-linux-x86-64.so.2 (0x00007fce1fc90000)
libtinfo.so.5 => /lib64/libtinfo.so.5 (0x00007fce12289000)

I attach here the binaries sddssort and example.

Code: Select all

sddssort -nondominate -col=DAPenalty -col=LTPenalty -col=ChromPenalty lattice_V0316_28s16o2.all myfile.sort ; sddsprintout -col=Rank myfile.sortsddssort -nondominate -col=DAPenalty -col=LTPenalty -col=ChromPenalty lattice_V0316_28s16o2.all myfile.sort ; sddsprintout -col=Rank myfile.sort

error: unknown switch: nondominate
Error (sddsprintout): ?
Printout for SDDS file myfile.sort

    Rank
------------
          1
          0
          1
          0
          1
          0
          1
          0
          1
          0
          1
          0
          2
          0
          2
          0
          2
          0
          2
          0
          3
          0
          3
          0
          3
          0
          3
          0
          3
          0
          3
          0
          3
          0
          3
          0
          4
          0
          4
          0

[soliday]

You are running three commands back to back, sddssort, sddsprintout and sddsprintout. The error you posted is coming from the first call to sddsprintout. It does not have a -nondominate option. The printout that you are seeing is from the last call to sddsprintout.

[nadolski]

Sorry for the typo, the command we run is :

excerpt from geneticOptimizer
sddssort -nondominate -col=DAPenalty -col=LTPenalty -col=ChromPenalty lattice_V0316_28s16o2.all myfile.sort
sddsprintout -col=Rank myfile.sort

the sorting command works well with elegant 34.3.1, Jul 4 2018,
Ir does not work with elegant 2021.3.0, Sep 21 2021

Could you please check with the file lattice_V0316_28s16o2.all provided in the previous post?

best regards,

Laurent.

Expect results :
sddssort -nondominate -col=DAPenalty -col=LTPenalty -col=ChromPenalty lattice_V0316_28s16o2.all myfile.sort ; sddsprintout -col=Rank myfile.sort
Printout for SDDS file myfile.sort

Rank
------------
1
1
1
1
2
2
2
2
2

[soliday]

I am not familiar with how non-dominate sorting works. I will forward this question on to the programmer who added this feature.

[soliday]

There was a bug in sddssort.c due to a feature change in the SDDS format. It has been fixed. I am attaching the updated sddssort.c file since you said you are building from source code.

[nadolski]

Great !
we will test it today.
Please let us know when the RPM for REH7 and RH8 are ready for download for our other platforms.

[nadolski]

This works well. Thanks again for the short reply.

[soliday]

I am attaching a newer version of the file. Apparently there was an issue with the yesterday's version. I don't know the details.
As for RPMs. There will be a new release today. I have been working on it for about a week now. This change will not be in the prebuilt RPMs but I will put it into the source release so that RPMs built with the Build-AOP-RPMs script will get this update.

[blanco-garcia]

Dear Soliday, thank you very much !

o