[Beowulf] Re: [Bioclusters] Direct connect infiniband/quadrics?
gotero at linuxprophet.com
Tue Mar 15 10:10:32 PST 2005
Speaking to the workable question--I've built small clusters just like
you described with Quadrics, Infiniband and Rocks. Quadrics support
isn't built into Rocks, but the Quadrics folks typically make their
software Rocks-aware. Infiniband support is supposedly built into
Rocks, but I haven't heard any success stories with it. It seems that
it's too Infinicon-centric. So you may be limited to using Infinicon
gear to make that work. I have experience making the openIB stuff work
with Rocks, so that is the path I would recommend. Mellanox and Topspin
are good to work with, as is Quadrics, when it comes to shoe-horning
their software onto Rocks clusters.
You'll see better performance even with two nodes. But I'd take a hard
look at what it will cost you to scale out with either interconnect and
decide if the difference in latency is worth the difference in price.
On Mar 15, 2005, at 9:22 AM, Farul Mohd Ghazali wrote:
> Has anyone had any experience with a direct connect/point-to-point
> implementation of Quadrics or Inifiniband? I talked to a small lab
> doing some computational chemistry and molecular dynamics work and
> they're interested in setting up a cluster but there is a need to
> justify the cost of a cluster before the budget can be approved.
> During the discussion, the idea of using direct connect infiniband or
> quadrics on two dual or quad Opteron nodes came up as a testbed
> platform to justify to management. From a price point of view, this is
> very attractive since it'll probably cost less than $40,000 (two quad
> Opterons, two Quadrics cards) for a testbed system. Money is tight...
> So, is this setup workable? In theory this should be faster than a
> gigabit based interconnect, even if it's just two nodes but I'd
> welcome any other ideas/suggestions. Thanks.
> -- "Leadership & Life-long Learning" --
> Farul Mohd. Ghazali
> Manager, Systems & Bioinformatics
> Open Source Systems Sdn. Bhd.
> www.aldrich.com.my Tel: +603-8656 0139/29 Fax: +603-8656 0132
> Bioclusters maillist - Bioclusters at bioinformatics.org
Glen Otero Ph.D.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 2264 bytes
Desc: not available
Url : http://www.scyld.com/pipermail/beowulf/attachments/20050315/4a571112/attachment.bin
More information about the Beowulf