Re: [OMPI devel] Average collective result graphs
Hello everyone, Please review the new collectives tuned defaults: https://github.com/open-mpi/ompi/pull/7730 as soon as possible. This patch will affect most users and adds significant performance improvements for collective operations. Thanks, William Zhang From: "Zhang, William" Date: Tuesday, May 12, 2020 at 4:57 PM To: Marco Atzeri via devel Subject: Re: Average collective result graphs Hello everyone, I have created a PR for the new collectives defaults. Please review this so it can be merged before the 5.0 branch. Thanks, William Zhang From: "Zhang, William" Date: Monday, May 11, 2020 at 1:06 AM To: Marco Atzeri via devel Subject: Average collective result graphs Hello everyone, I have aggregated and averaged the results from the everyone who submitted data for collective runs. You can view the graphs here if you’re interested: https://drive.google.com/open?id=1MV5E9gN-5tootoWoh62aoXmN0jiWiqh3 Algorithm 0 represents the current defaults. I will refer to these graphs when generating new defaults. Thanks, William Zhang
Re: [OMPI devel] Warning
Luis, With some low frequency we remove warnings from the code. In this particular instance the meaning of the code is correct, the ompi_info_t structure starts with an opal_info_t, but removing the warnings is good policy. In general we can either cast the ompi_info_t pointer directly to an opal_info_t pointer, or access the super field in the ompi_info_t structure (as it is done in ompi/mpi/c/win_create_dynamic.c). As in this instance we are explicitly using one of the MPI predefined info keys (without equivalent at the OPAL level) it is more clear if we cast it. George. On Fri, May 15, 2020 at 6:37 AM Luis via devel wrote: > Hi OMPI devs, > > I was wondewring if this warning is expected, if not, how should we > internally call ompi_win_create_dynamic? > > res = ompi_win_create_dynamic(MPI_INFO_NULL, comm, ); > ^ > In file included from pnbc_osc_internal.h:40, > from pnbc_osc_iallreduce.c:21: > ../../../../ompi/win/win.h:143:42: note: expected ‘opal_info_t *’ {aka > ‘struct opal_info_t *’} but argument is of type ‘struct ompi_info_t *’ > int ompi_win_create_dynamic(opal_info_t *info, ompi_communicator_t > *comm, ompi_win_t **newwin); > > > Regards, > Luis > The University of Edinburgh is a charitable body, registered in Scotland, > with registration number SC005336. >
Re: [OMPI devel] Warning
Luis, you can do this: struct ompi_info_t * info = _mpi_info_null.info; ret = ompi_win_create_dynamic(_null->super, comm, win); Cheers, Gilles On Fri, May 15, 2020 at 7:38 PM Luis via devel wrote: > > Hi OMPI devs, > > I was wondewring if this warning is expected, if not, how should we > internally call ompi_win_create_dynamic? > > res = ompi_win_create_dynamic(MPI_INFO_NULL, comm, ); > ^ > In file included from pnbc_osc_internal.h:40, > from pnbc_osc_iallreduce.c:21: > ../../../../ompi/win/win.h:143:42: note: expected ‘opal_info_t *’ {aka > ‘struct opal_info_t *’} but argument is of type ‘struct ompi_info_t *’ > int ompi_win_create_dynamic(opal_info_t *info, ompi_communicator_t > *comm, ompi_win_t **newwin); > > > Regards, > Luis > The University of Edinburgh is a charitable body, registered in Scotland, > with registration number SC005336.
[OMPI devel] Warning
Hi OMPI devs, I was wondewring if this warning is expected, if not, how should we internally call ompi_win_create_dynamic? res = ompi_win_create_dynamic(MPI_INFO_NULL, comm, ); ^ In file included from pnbc_osc_internal.h:40, from pnbc_osc_iallreduce.c:21: ../../../../ompi/win/win.h:143:42: note: expected ‘opal_info_t *’ {aka ‘struct opal_info_t *’} but argument is of type ‘struct ompi_info_t *’ int ompi_win_create_dynamic(opal_info_t *info, ompi_communicator_t *comm, ompi_win_t **newwin); Regards, Luis The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. pEpkey.asc Description: application/pgp-keys