[Trilinos-Users] Epetra_CrsMatrix FillComplete error

홍광우 yuriprime at kaist.ac.kr
Tue Jun 24 22:58:08 MDT 2014


Thank you for your response,


I checked the return code of all Epetra functions, but there is no error return code when I construct matrices.


The program runs well with small size matrices, but gives error with large size matrices.


For example, the error occurs when I use large matrices which needs about 60% of total memory. However, when I use small matrices which needs about 5%, the program runs.


So, I think that the size of matrices is related to the error of FillComplete function.


Do you have any opinion about it?


Kwangwoo Hong
 


-----------------------원본메일-----------------------
보낸사람 : "Heroux, Mike" <MHeroux at CSBSJU.EDU>
받는사람: 홍광우<yuriprime at kaist.ac.kr>,"trilinos-users at software.sandia.gov" <trilinos-users at software.sandia.gov>
보낸날짜: 2014-06-25 13:25:01 GMT +0900 (Asia/Seoul)
제목: Re: [Trilinos-Users] Epetra_CrsMatrix FillComplete error

Kwangwoo,



If you are not already doing so, please check all return values from Epetra functions.  It appears you are stumbling on some error code and there may be evidence of what is wrong.



Mike



From: 홍광우 <yuriprime at kaist.ac.kr<mailto:yuriprime at kaist.ac.kr>>

Date: Tuesday, June 24, 2014 10:13 AM

To: "trilinos-users at software.sandia.gov<mailto:trilinos-users at software.sandia.gov>" <trilinos-users at software.sandia.gov<mailto:trilinos-users at software.sandia.gov>>

Subject: [Trilinos-Users] Epetra_CrsMatrix FillComplete error





Dear Trilinos developers and users,





Sorry for the repeated question. This time, I have a question about FillComplete of Epetra_CrsMatrix.





When I call FillComplete, the following error occurs and the program killed.





terminate called after throwing an instance of 'int'



[r1i0n10:26467] *** Process received signal ***



[r1i0n10:26467] Signal: Aborted (6)



[r1i0n10:26467] Signal code:  (-6)



[r1i0n10:26467] [ 0] /lib64/libpthread.so.0(+0xf4c0) [0x7f28e6b654c0]



[r1i0n10:26467] [ 1] /lib64/libc.so.6(gsignal+0x35) [0x7f28e4d279a5]



[r1i0n10:26467] [ 2] /lib64/libc.so.6(abort+0x175) [0x7f28e4d29185]



[r1i0n10:26467] [ 3] /usr/lib64/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x12d) [0x7f28e6615aad]



[r1i0n10:26467] [ 4] /usr/lib64/libstdc++.so.6(+0xbcc36) [0x7f28e6613c36]



[r1i0n10:26467] [ 5] /usr/lib64/libstdc++.so.6(+0xbcc63) [0x7f28e6613c63]



[r1i0n10:26467] [ 6] /usr/lib64/libstdc++.so.6(+0xbcd5e) [0x7f28e6613d5e]



[r1i0n10:26467] [ 7] /appl/trilinos/kwangwoo/libs/lib/libepetra.so.11(_ZN15Epetra_CrsGraph15OptimizeStorageEv+0x655) [0x7f28ea05fe55]



[r1i0n10:26467] [ 8] /appl/trilinos/kwangwoo/libs/lib/libepetra.so.11(_ZN16Epetra_CrsMatrix15OptimizeStorageEv+0xc1) [0x7f28ea08a241]



[r1i0n10:26467] [ 9] /appl/trilinos/kwangwoo/libs/lib/libepetra.so.11(_ZN16Epetra_CrsMatrix12FillCompleteERK10Epetra_MapS2_b+0xf7) [0x7f28ea0899c7]



[r1i0n10:26467] [10] /appl/trilinos/kwangwoo/libs/lib/libepetra.so.11(_ZN16Epetra_CrsMatrix12FillCompleteEb+0x1a) [0x7f28ea08982a]



[r1i0n10:26467] [11] exe(_ZN16Core_Hamiltonian16construct_matrixEv+0x7a3) [0x526153]



[r1i0n10:26467] [12] exe(_ZN16Core_HamiltonianC1EP5BasisP12Grid_SettingP5AtomsP10Epetra_MapPN7Teuchos13ParameterListE+0x55) [0x525965]



[r1i0n10:26467] [13] exe(_ZN2SpC1EP5BasisP12Grid_SettingP5AtomsP5GuessP10Epetra_MapPN7Teuchos13ParameterListE+0x1d4) [0x53eb64]



[r1i0n10:26467] [14] exe(main+0x212e) [0x58913e]



[r1i0n10:26467] [15] /lib64/libc.so.6(__libc_start_main+0xfd) [0x7f28e4d13c5d]



[r1i0n10:26467] [16] exe() [0x439d49]



[r1i0n10:26467] *** End of error message ***



Aborted





I don't know what the solution is. Does anyone know about the above problem?





Sincerely,





Kwangwoo Hong





Kwangwoo Hong



Ph.D Candidate



Computational Quantum Molecular Science Lab.



Department of Chemistry, KAIST



(Tel) +82-42-350-2855











-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://software.sandia.gov/pipermail/trilinos-users/attachments/20140625/333d5366/attachment.html>


More information about the Trilinos-Users mailing list