Re: [OMPI devel] knem support in sm btl

Kewl - just wanted to double check as this had come up before, and people felt strongly about it. I didn’t see any clear way to pass the error back to MPI_Init, which is why I started this thread. Sounds like one doesn’t exist, so exit is probably the only real option. Thanks Ralph > On Dec

Re: [OMPI devel] knem support in sm btl

I was not advocating calling exit. I was merely suggesting that due to earliness in the initialization process, and to the fact that we are lacking the infrastructure to abort because a specific user request cannot be complied to, calling exit seems like a reasonable bandaid. George. On Fri,

Re: [OMPI devel] knem support in sm btl

Let me get this straight - you are advocating that I call “exit” directly from within a library?? I thought that was “verboten” - MPI_Init should just return an error somehow, yes? > On Dec 4, 2014, at 12:35 PM, Jeff Squyres (jsquyres) > wrote: > > Oh, good catch --

Re: [OMPI devel] knem support in sm btl

Oh, good catch -- thanks. I wouldn't call abort -- that will dump core. Just show_help() and exit(nonzero), I guess. On Dec 4, 2014, at 3:31 PM, George Bosilca wrote: > You can't use the PML error reporting mechanism in this particular instance, > it is too early in

Re: [OMPI devel] knem support in sm btl

You can't use the PML error reporting mechanism in this particular instance, it is too early in the setup process (in the BTL component init function) and the PML has not setup the error callback yet. This function is called during the MPI_Init, at a time where most of the Open MPI infrastructure

Re: [OMPI devel] knem support in sm btl

You're supposed to call the PML error handler, which was passed down to the BTL during initialization. That is, the BTL registers a btl_register_error function with the PML. The PML then calls this function and passes in its error handler function pointer. The BTL can then use that error

Re: [OMPI devel] RTLD_GLOBAL question

2014-12-04 17:29 GMT+06:00 Jeff Squyres (jsquyres) : > On Dec 3, 2014, at 11:35 PM, Artem Polyakov wrote: > > > Jeff, I must admit that I don't completely understand how your fix work. > Can you explan me why this veriant was failing: > > > >

Re: [OMPI devel] RTLD_GLOBAL question

On Dec 3, 2014, at 11:35 PM, Artem Polyakov wrote: > Jeff, I must admit that I don't completely understand how your fix work. Can > you explan me why this veriant was failing: > > CPPFLAGS="-I$srcdir/opal/libltdl/" > AC_EGREP_HEADER([lt_dladvise_init],