Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: Re: [OMPI devel] alps ras patch for SLURM
From: Matney Sr, Kenneth D. (matneykdsr_at_[hidden])
Date: 2010-07-12 08:40:51

Hi Ralph,

I think that it would be overstating the case to say that I am re-assuming
those duties. Rather, I am trying to fill the gap in a minimal sense while
we locate a replacement for Rainer. I expect to help our replacement
get up to speed on portals and ALPS; but, I have too many other duties
to be able to provide full OMPI support. Here is a plug for my daytime job:

I do the Lustre User Toolkit and IOTA. The web pages will be improved
before too long. The current state reflects my preference for development
over writing web pages. In addition, new versions of these are in development
as well as a new parallel tool and another tool for use on Lustre. Other
members, of the group I work in, are developing additional tools for Lustre.
The Lustre User Toolkit is compatible with OMPI and runs on our research cluster,
as well as the Cray XTs.

At the same time, I do not want OMPI to be unsupported on the Cray XT.
There may be a need for ORNL down the road.

I agree that, in the OMPI context, the "slurm" comment generates confusion.
The source of the environmental variable is immaterial to OMPI. If you feel
that I should do the patch for Jerome, just say so. Thanks.
-- Ken

On Jul 10, 2010, at 10:37 AM, Ralph Castain wrote:

> Sounds good then.
> I only got into this thread because (a) the reference to slurm, and (b) with Rainer's departure, I wasn't sure if someone else was going to pickup the alps support. Since you are re-assuming those latter duties (yes?), and since this actually has nothing to do with slurm itself, I'll let you decide when/if to deal with the patch.
> I would only suggest that you remove the "slurm" comment from it as it is definitely confusing.
> Thanks
> Ralph


> _______________________________________________
> devel mailing list
> devel_at_[hidden]