Yes. Also from talking with Scott they're also looking for help to evaluate the hardware. Hence my note that I'd be able to lend a hand in baseline benchmarking (IOR, iperf, etc), evaluating how provisioning works, etc. If we want to run application benchmarks on it, I'd want some time from Kumar's folks. Assuming someone on his team is interested. Scott, Dan and Ryan are interested is evaluating it for OpenStack use. Ben On Aug 20, 2014, at 11:27 AM, Rajeev Thakur <[email protected]> wrote:
Are they planning to get it and just want JLSE to host it?
Rajeev
On Aug 19, 2014, at 5:30 PM, "Allen, Benjamin S." <[email protected]> wrote:
I'm interested in hosting this, and have chatted with Scott about the eval. I can help out with installing, evaluating the systems administration side of the hardware, provisioning, baseline benchmarking, etc.
Ben
On Aug 19, 2014, at 5:15 PM, Kalyan Kumaran <[email protected]> wrote:
Begin forwarded message:
From: "Murphy-Olson, Daniel E." <[email protected]> Subject: Seamicro Evaluation Date: August 19, 2014 at 5:08:28 PM CDT To: "Thakur, Rajeev" <[email protected]>, "Kumaran, Kalyanasundaram" <[email protected]>
Hi Rajeev and Kumar,
The Magellan infrastructure team, Scott Devoid, Ryan Aydelott and I have been looking into several different hardware options that could be used as building blocks for Magellan's next refresh cycle. One interesting option is from Seamicro, currently part of AMD. We're interested in working with JLSE to evaluate the Seamicro hardware.
Here is some more information regarding the machine we're interested in: http://www.seamicro.com/sm15000 http://www.seamicro.com/sites/default/files/SM_TO02_64_v2%205.pdf
The primary areas we'd like to evaluate are: Internal torus network scalability How the system deals with changes in storage and network utilization on the fabric How well the system integrates with the Openstack control plane Serviceability of the components when operating under Openstack Compute performance under Openstack
While we're primarily interested in the cloud computing capabilities, this platform may have HPC capabilities as well.
Currently we have an NDA addendum we've received from AMD for this, but we don't yet have an evaluation agreement.
Would you be interested in working with us on this? If so, how do we move forward?
Thanks, Dan
_______________________________________________ Admins mailing list [email protected] https://lists.jlse.anl.gov/mailman/listinfo/admins
_______________________________________________ Admins mailing list [email protected] https://lists.jlse.anl.gov/mailman/listinfo/admins