Hi everybody,
I have 6 desktop computers with the following specifications (all of them are the same):
CPU: 2 GHz, 2 cores, 8 Hyper Threading
RAM: 12 Gb
HDD: 500 Gb
Windows 7 (which I need to change to Windows Server with HPC pack)
I need to connect them together and make a cluster for abaqus analysis. I'm thinking about using Microsoft HPC pack since it seems to be an easy and user friendly option (I don't have any experience in that area).
Since I already have the computers, I only need to spend on the softwares (OS, HPC pack, etc) and some network and cable stuff. My budget is limited to $1000 but I can push it to $2000 if it's necessary. <!--break-->
First I need to know if it worth even trying to make a 12 core cluster or not? How good Abaqus can handle 48 threads? I know Abaqus is good at multi-threading but this won't be a 48 core cluster, it will be a 12 core with the ability of having 48 threads so that's why I'm a little hesitant.
Second I already Googled the procedure of making a cluster from scratch but I couldn't find anything that explains it in a non-technical language. Could anyone (preferably whoever done this before) explain from where I should start? Like do I need to buy Windows Server for all the machines or just one license for the head node is enough?
Any recommendation would be greatly appreciated.
Thank you,
Maysam
ps. I'm performing fatigue and fracture analysis using abaqus.
Normal
0
false
false
false
EN-US
X-NONE
X-NONE
/* Style Definitions */
table.MsoNormalTable
{mso-style-name:"Table Normal";
mso-tstyle-rowband-size:0;
mso-tstyle-colband-size:0;
mso-style-noshow:yes;
mso-style-priority:99;
mso-style-parent:"";
mso-padding-alt:0in 5.4pt 0in 5.4pt;
mso-para-margin-top:0in;
mso-para-margin-right:0in;
mso-para-margin-bottom:10.0pt;
mso-para-margin-left:0in;
line-height:115%;
mso-pagination:widow-orphan;
font-size:11.0pt;
font-family:"Calibri","sans-serif";
mso-ascii-font-family:Calibri;
mso-ascii-theme-font:minor-latin;
mso-hansi-font-family:Calibri;
mso-hansi-theme-font:minor-latin;}
The “Michigan Nano Computational Cluster” (MNC2) consists of 14 nodes.
Each compute node contains a hexa-core Intel Xeon X5660 processor
running at 2.66 GHz, 24 GB of memory. MNC2 contains 168 processing cores
linked with Gigabit ethernet, 12 TB of disk space.
MNC2 currently hosts simulation tools for micro/nanoscale systems
including codes for first principles calculation, photonic devices,
molecular dynamics, multiphysics, and multiscale modeling.
The Cluster is free for use. Researchers interested in using the cluster,
Please contact Dr Behrouz Shiari, bshiari@umich.edu or visit http://lnf.umich.edu/nnin-at-michigan/index.php/computation/getting-star...
National Nanotechnology Infrastructure Network Computation (NNIN/C)
Project at Michigan provides free computation cycling time for a limited
time (July 25, 2012 to Dec. 20, 2012) to support researchers in
Nanotechnology (Nano/Microsystems, Nanomaterials, Multiscale modeling,
etc). To get a free computation account, please go to:
www.lnf.umich.edu/NNIN/Computation/Registration/index.aspx
For more information, please contact bshiari@umich.edu
or:
National Nanotechnology Infrastructure Network Computation (NNIN/C) Project at Michigan provides free computation cycling time for a limited time (Sept. 15, 2011 to Jan. 15, 2012) to support researches in Nanotechnology (Nano/Microsystems, Nanomaterials, Multiscale modeling, etc). To get a free computation account, please go to:
www.lnf.umich.edu/NNIN/Computation/Registration/index.aspx
For more information, please contact bshiari@umich.edu
or:
June 18 Columbus, OH
Guest Speaker: Dan Turk,Honda R&D Americas\http://www.simulia.com/events/hpc/columbus_hpc_6-18.html
June 24 Minneapolis; MN
Guest Speaker; Paul Radavic, Design Analysis Manager, Mercury Marine
http://www.simulia.com/events/hpc/minneapolis_hpc_6-24.html
June 30 Cleveland, OH
Guest Speaker: To be announced
http://www.simulia.com/events/hpc/cleveland_hpc_6-30_simweb.html
By attending this free one-day seminar, you will discover the answers to common questions surrounding computing clusters. You will be provided guidance on how to specify, install, and maintain an affordable, computing cluster solution.
Plus, learn about the latest advancements in parallel computing performance with Abaqus software, and how you can leverage an affordable computing cluster to dramatically improve simulation throughput.
For more information and to register visit: www.simulia.com