Hi Neks, I was wondering if someone could tell me the implications of commenting this section of code out in Nekbone: if (nval.lt.(10000+max(lp,lelg))) then if(nid.eq.0) then write(6,*) 'ABORT: MPI_TAG_UB too small!' endif call exitt endif We are testing Nekbone with an experimental MPI library. Regards, Ken The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
To answer your question I need to know why you want to comment this check? Am 8/18/16, 7:02 AM schrieb "[email protected] on behalf of [email protected]" unter <[email protected] on behalf of [email protected]>:
Hi Neks,
I was wondering if someone could tell me the implications of commenting this section of code out in Nekbone:
if (nval.lt.(10000+max(lp,lelg))) then if(nid.eq.0) then write(6,*) 'ABORT: MPI_TAG_UB too small!' endif call exitt endif
We are testing Nekbone with an experimental MPI library.
Regards, Ken
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
Basically, the MPI library that we are testing sets MPI_TAG_UB=65534, which is quite low to be used with higher numbers of lp. On 18/08/16 19:33, [email protected] wrote:
To answer your question I need to know why you want to comment this check?
Am 8/18/16, 7:02 AM schrieb "[email protected] on behalf of [email protected]" unter <[email protected] on behalf of [email protected]>:
Hi Neks,
I was wondering if someone could tell me the implications of commenting this section of code out in Nekbone:
if (nval.lt.(10000+max(lp,lelg))) then if(nid.eq.0) then write(6,*) 'ABORT: MPI_TAG_UB too small!' endif call exitt endif
We are testing Nekbone with an experimental MPI library.
Regards, Ken
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
_______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
I checked the source (nekbone 3.1) and it looks like the fortran part does not use any tagged messages but the C part does. In fact, there are C parts using the MPI rank to tag messages. So I am afraid commenting the check won¹t help. Cheers, Stefan Am 8/18/16, 8:55 AM schrieb "[email protected] on behalf of [email protected]" unter <[email protected] on behalf of [email protected]>:
Basically, the MPI library that we are testing sets MPI_TAG_UB=65534, which is quite low to be used with higher numbers of lp.
On 18/08/16 19:33, [email protected] wrote:
To answer your question I need to know why you want to comment this check?
Am 8/18/16, 7:02 AM schrieb "[email protected] on behalf of [email protected]" unter <[email protected] on behalf of [email protected]>:
Hi Neks,
I was wondering if someone could tell me the implications of commenting this section of code out in Nekbone:
if (nval.lt.(10000+max(lp,lelg))) then if(nid.eq.0) then write(6,*) 'ABORT: MPI_TAG_UB too small!' endif call exitt endif
We are testing Nekbone with an experimental MPI library.
Regards, Ken
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
_______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ Nek5000-users mailing list [email protected] https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
participants (1)
-
nek5000-users＠lists.mcs.anl.gov