Re: [OMPI devel] heterogeneous OpenFabrics adapters

2008-05-13 Thread Jeff Squyres
On May 13, 2008, at 12:24 PM, Jon Mason wrote: - Because the BSRQ info is on the component (i.e., global), we should detect when multiple different receive_queues values are specified and gracefully abort. How would we verify that the remote receive_queues values are the same? By passing

Re: [OMPI devel] heterogeneous OpenFabrics adapters

2008-05-13 Thread Jon Mason
On Monday 12 May 2008 07:37:54 pm Jeff Squyres wrote: > Short version: > -- > > I propose that we should disallow multiple different > mca_btl_openib_receive_queues values (or receive_queues values from > the INI file) to be used in a single MPI job for the v1.3 series. > > More details

Re: [OMPI devel] heterogeneous OpenFabrics adapters

2008-05-13 Thread Pavel Shamis (Pasha)
Jeff, Your proposal for 1.3 sounds ok for me. For 1.4 we need review this point again. The qp information is split over 3 different structs: mca_btl_openub_module_qp_t (used by module), mca_btl_openib_qp_t (used by component) and mca_btl_openib_endpoint_qp_t (used by endpoint). Need see how we

Re: [OMPI devel] heterogeneous OpenFabrics adapters

2008-05-12 Thread Jeff Squyres
Short version: -- I propose that we should disallow multiple different mca_btl_openib_receive_queues values (or receive_queues values from the INI file) to be used in a single MPI job for the v1.3 series. More details: - The reason I'm looking into this heterogeneit

Re: [OMPI devel] heterogeneous OpenFabrics adapters

2008-05-12 Thread Jeff Squyres
After looking at the code a bit, I realized that I completely forgot that the INI file was invented to solve at least the heterogeneous- adapters-in-a-host problem. So I amended the ticket to reflect that that problem is already solved. The other part is not, though -- consider two MPI proc