Re: [myrinet] RE: Myrinet Socket Migration?
"Jared Hodge" <[email protected]>
| Newsgroups | gmane.network.myrinet.general |
|---|---|
| Message-ID | <[email protected]> |
I guess my question wasn't as clear as I thought it was. What I meant by socket migration is not going to a different platform, but stopping a running job and starting it again on a different machine. In other words, checkpointing. Does anyone have any experience with this? "Cliff, Gary" wrote: > > Hi, > > I'm running Myrinet on Solaris and VxWorks and build both of them locally. > I belive the hardest part of the migration process would be to understand > the socket infrastructure of the platform you're porting to. Once you > understand how the hooks work, the rest would be easy. Usually, there's > example code to work from. > > As a side note, I found it very easy to port MPICH-GM to VxWorks. > > Gary Cliff > Senior Software Engineer > Computing Devices Canada Ltd > Email : [email protected] > Voice : (613) 596-7692 > Fax : (613) 596-7125 > > -----Original Message----- > From: Jared Hodge [mailto:[email protected]] > Sent: Thursday, August 16, 2001 11:54 AM > To: [email protected] > Subject: [myrinet] Myrinet Socket Migration? > > I was wondering if anyone on this mailing list has any insight as to > how difficult implementing Myrinet socket migration would be (so that > MPICH-GM jobs could be migrated to different nodes). I've seen recent > discussion on the Beowulf mailing list about TCP/IP migration > difficulty, but was wondering if Myrinet might be any easier. I'm not > very familiar with the Myrinet drivers, which is why I'm asking people > who are. Thanks. > -- > Jared Hodge > Institute for Advanced Technology > The University of Texas at Austin > 3925 W. Braker Lane, Suite 400 > Austin, Texas 78759 > > Phone: 512-232-4460 > Fax: 512-471-9096 > Email: [email protected] -- Jared Hodge Institute for Advanced Technology The University of Texas at Austin 3925 W. Braker Lane, Suite 400 Austin, Texas 78759 Phone: 512-232-4460 Fax: 512-471-9096 Email: [email protected]