[Omp] segmentation fault with private variable
Dieter an Mey
anmey at rz.rwth-aachen.de
Thu Feb 1 05:25:53 PST 2007
You probably need to increase the stacksize for the master thread
(ulimit -s ...) and also for the slave threads, which is platform-dependent.
Check the compiler docu.
best regards
Dieter
Julien Taillard schrieb:
> Dear all,
>
> I have a segmentation fault when I execute this code with more than 1
> thread with xlf_90 compiler.
> Is there any restriction in data size for private variable in OpenMP or
> have you got any idea of the problem ?
>
> If I remove m2Inst, it works fine...
>
> PROGRAM testInitTab
>
> INTEGER :: p0
> INTEGER :: OMP_GET_THREAD_NUM
> INTEGER , DIMENSION(0:999 , 0:999) :: matrix1Inst
> INTEGER , DIMENSION(0:999 , 0:999) :: matrix2Inst
> INTEGER , DIMENSION(0:999 , 0:999) :: Cinst
>
> INTEGER , DIMENSION(0:499 , 0:999) :: inAInst
> INTEGER, DIMENSION(0:999,0:499) :: inBInst
> INTEGER, DIMENSION(0:499,0:499) :: m2Inst
>
> !$OMP PARALLEL PRIVATE(p0,inAInst,inBInst,m2Inst)
> SHARED(matrix1Inst ,matrix2Inst ,Cinst)
>
> p0=OMP_GET_THREAD_NUM()
>
> PRINT *, "Hello from ",p0
>
> !$OMP BARRIER
>
> inAInst = 0
> inBInst = 0
> m2Inst =0
>
> !$OMP END PARALLEL
> END PROGRAM testInitTab
>
>
>
> Best regards,
> Julien
> _______________________________________________
> Omp mailing list
> Omp at openmp.org
> http://openmp.org/mailman/listinfo/omp
>
--
--------------------------------------------------------------------
Dieter an Mey
High Performance Computing Hochleistungsrechnen
RWTH Aachen University Rechen- und Kommunikations-
Center for Computing and Communication zentrum der RWTH Aachen
phone: ++49-(0)241-80-24377 Seffenter Weg 23
fax: ++49-(0)241-80-22134 52074 Aachen, Germany
email: anmey at rz.rwth-aachen.de
--------------------------------------------------------------------
More information about the Omp
mailing list