[Beowulf] channel bonding
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
steve_heaton at iinet.net.au steve_heaton at iinet.net.auSun Aug 7 16:39:51 PDT 2005
- Previous message: [Beowulf] Begginer questions
- Next message: [Beowulf] Re: Beowulf Digest, Vol 18, Issue 9
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
G'day Andrew and all Not sure if MP_Lite is what you were thinking of? Sorry if this turns out to be noise. http://www.scl.ameslab.gov/Projects/MP_Lite/ Cheers Steve > From: Andrew Piskorski <atp at piskorski.com> > Subject: Re: [Beowulf] channel bonding > To: beowulf at beowulf.org > Cc: Tim Mattox <tmattox at gmail.com>, asada at dsee.fee.unicamp.br > Message-ID: <20050807164056.GA18738 at piskorski.com> > Content-Type: text/plain; charset=us-ascii > <snip> > > The current latest is from 21 June 2005: > http://unc.dl.sourceforge.net/sourceforge/bonding/bonding.txt > > That's the stock channel bonding shipped witht the kernel. I vaguely > remember reading (probably a year or more ago) that its maximum > througput is achieved with 2 gigabit cards - adding a 3rd card > actually decreases througput. (Is this still true?) > > Also, that there's a different implementation somewhere, not shipped > with the kernel, which is both faster with two cards, and allows > further bandwith increase by adding a 3rd bonded gigabit ethernet > card. Unfortunately, I can't remember what that project is called, > nor where I read about it. Does anyone else remember? > > -- > Andrew Piskorski <atp at piskorski.com> > http://www.piskorski.com/
- Previous message: [Beowulf] Begginer questions
- Next message: [Beowulf] Re: Beowulf Digest, Vol 18, Issue 9
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
