Thanks for your question.
That's because many Big Data middleware are using Java/Scala
currently. We are hoping our Java-based library can be easily used by
the existing works in the Big Data community.
We understand the JVM overhead but it also brings the benefits from
the productivity angle. In the future, we have the plan to release the
C implementation.
Thanks
DataMPI Team
On Sun, Aug 17, 2014 at 9:46 PM, <
manue...@gmail.com> wrote:
> I think the jvm with the garbage collection and other overheads prevents higher performance
>
> --
> You received this message because you are subscribed to the Google Groups "MPI-D" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
mpi-d+un...@googlegroups.com.
> For more options, visit
https://groups.google.com/d/optout.