# Segmentation fault when running Abinito Relax

2 posts / 0 new
Segmentation fault when running Abinito Relax
#1

Hi everyone,
I try to using AbinitioRelax of Rosetta3.4 to predict the structure of a amino acid sequence (308 residues).
However, even when I used the simple flags file (with no option for optimal performace) that just required the fasta and fragments input files, the script couldn't be run.
Error was "Segmentation fault 11"
First I thought my input file had something wrong, then, I tried to run the demo script (rosetta3.4/rosetta_demos/abinitio) with full inputs file and the flags file like below:

-in:file:fasta ./input_files/1elwA.fasta
-in:file:frag3 ./input_files/aa1elwA03_05.200_v1_3
-in:file:frag9 ./input_files/aa1elwA09_05.200_v1_3
-database /home/hongtham/SW/rosetta3.4/rosetta_database
-in:file:native ./input_files/1elw.pdb
-abinitio:relax
-nstruct 1
-out:pdb
-use_filters true
-psipred_ss2 ./input_files/1elwA.psipred_ss2
-abinitio::increase_cycles 10
-abinitio::rg_reweight 0.5
-abinitio::rsd_wt_helix 0.5
-abinitio::rsd_wt_loop 0.5
-relax::fast

and the executive script was:
RosettaHOME=/home/hongtham/SW/rosetta3.4/rosetta_source/bin
/usr/local/bin/mpiexec -np $NP$RosettaHOME/AbinitioRelax.mpi.linuxgccrelease @flags > log

However, when the abinitiorelax run to " Stage 2 Folding with score1 for 20000", then it stopped suddenly. The error file was:

*** Process received signal ***
Signal: Segmentation fault (11)
Signal code: Address not mapped (1)
Failing at address: 0x58
[ 0] /lib64/libpthread.so.0 [0x349480ebe0]
[ 1] /usr/lib64/libstdc++.so.6(_ZNSs6assignERKSs+0x1e) [0x3499c9ca2e]
[ 2] /home/hongtham/SW/rosetta3.4/rosetta_source/build/src/release/linux/2.6/64/x86/gcc/4.1/mpi/libprotocols.1.so(_ZN9protocols7jobdist18BaseJobDistributor8next_jobERN7utility7pointer10owning_ptrINS0_8BasicJobEEERi+0x25d) [0x2b42601c4c8d]
[ 3] /home/hongtham/SW/rosetta3.4/rosetta_source/build/src/release/linux/2.6/64/x86/gcc/4.1/mpi/libprotocols_b.5.so(_ZN9protocols8abinitio18AbrelaxApplication4foldERN4core4pose4PoseEN7utility7pointer10owning_ptrINS0_8ProtocolEEE+0x1bfc) [0x2b425c3bd3fc]
[ 4] /home/hongtham/SW/rosetta3.4/rosetta_source/build/src/release/linux/2.6/64/x86/gcc/4.1/mpi/libprotocols_b.5.so(_ZN9protocols8abinitio18AbrelaxApplication3runEv+0x507) [0x2b425c3c7087]
[ 5] /home/hongtham/SW/rosetta3.4/rosetta_source/bin/AbinitioRelax.mpi.linuxgccrelease(main+0xc6) [0x408ec6]
[ 6] /lib64/libc.so.6(__libc_start_main+0xf4) [0x349381d994]
[ 7] /home/hongtham/SW/rosetta3.4/rosetta_source/bin/AbinitioRelax.mpi.linuxgccrelease [0x408d29]
*** End of error message ***
--------------------------------------------------------------------------
mpiexec noticed that process rank 6 with PID 1468 on node node02.mark.pknu.ac.kr exited on signal 11 (Segmentation fault).

I have checked all the input files and the executive script. It was copied from directory rosetta_demos exactly. Unfortunately, I could'n find any solution. Can anybody can help me?
Thank you so much.

Hong Tham

Post Situation:
Mon, 2014-12-15 07:16
Hongtham

It looks like it has to do specifically with you running things under MPI. What MPI library are you using? Do you get similar errors if you run with the non-MPI compiled application?

Also, Rosetta3.4 is now several years old at this point. You may want to try updating to the latest weekly release, and seeing if the problem has been addressed since.

Finally, a 308 aa protein is likely going to be too large for straight abinitio modeling. I'd highly recommend looking for homologs with which you can do homology modeling, or, if possible, experimental information you could use to constrain your system.

Fri, 2015-01-02 11:45
rmoretti