Le 18/07/2013 13:23, Dave Love a écrit : > Mark Dixon <m.c.di...@leeds.ac.uk> writes: > >> On Mon, 15 Jul 2013, Elken, Tom wrote: >> ... >>> Hope these anecdotes are relevant to Open MPI users considering knem. >> ... >> >> Brilliantly useful, thanks! It certainly looks like it may be greatly >> significant for some applications. Worth looking into. > For general interest: there were some in the INRIA publications list, > and I think possibly from UTK. The papers also reference the previous, > less general, implementations (e.g. in MX as well as kcopy -- not that > I'm aware of kcopy making much difference to things running here). > > Maybe Brice can comment some time (post-vacances?).
Right, UTK did quite a lot of work on optimizing collectives over KNEM and looking at the impact on several applications. Have a look at the bottom of http://runtime.bordeaux.inria.fr/knem/ for references. On the Inria side, we mostly focused on KNEM itself, the main publication about this lower-level side of the software stack is the 2013 JPDC article (there's a link in the above URL too). Brice