Upload
jayson-johns
View
214
Download
0
Tags:
Embed Size (px)
Citation preview
Big Red II & Supporting InfrastructureCraig A. Stewart, Matthew R. Link, David Y
Hancock
Presented at IUPUI Faculty Council Information
Technology Subcommittee meeting 9 April 2013
Research Technologies
UITS
Indiana University
License TermsPlease cite as: Stewart, C.A. Big Red II & Supporting Infrastructure, 2013. Presentation. Presented at: IUPUI Faculty Council Information Technology Subcommittee meeting (Indianapolis, IN 9 April 2013). http://hdl.handle.net/2022/16981Items indicated with a © are under copyright and used here with permission. Such items may not be reused without permission from the holder of copyright except where license terms noted on a slide permit reuse. Except where otherwise noted, contents of this presentation are copyright 2013 by the Trustees of Indiana University. This document is released under the Creative Commons Attribution 3.0 Unported license (http://creativecommons.org/licenses/by/3.0/). This license includes the following terms: You are free to share – to copy, distribute and transmit the work and to remix – to adapt the work under the following conditions: attribution – you must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). For any reuse or distribution, you must make clear to others the license terms of this work.
Overview
• The big goal: transform the way IU generally uses advanced computing to advance basic research, scholarship, translational research, and artistic expression
• Big Red II• IU Bloomington Data Center• IU Research Network• Data Capacitor II• Research Home Directories• Quarry is still here!!!• IU Cyberinfrastructure Gateway
Flikr user hassanrafeek CC license
Big Red II – System Specifications
5
System Size – 1056 nodes (264 blades)• 344 XE6 compute nodes (86 blades)
2.5 GHz 16-core Abu Dhabi
64 GB system memory• 676 XK7 Compute (169 blades)
2.3 GHz 16-core Interlagos
NVIDIA K20 GPU (Kepler)32 GB system memory5 GB video memory
• 36 service & IO nodesFile System & Storage
• Boot Raid• 180 TB Lustre File system (1 GB/s)
Interconnect: GeminiTopology: 11 x 6 x 8 3D
TorusPeak Performance: 1.0003 PFLOPSTotal Memory Size: 43.6 TBTotal XE6 Cores: 11,008Total XK7 Cores: 10,816Total x86-64 Cores: 21,824
Big Red II – Can I log in yet?
8
• Stability testing & early user mode – April 2013• Dedication – April 26, 2013• General availability – sometime after that
Big Red II – What will we do with it?
9
• HPC applications in Extreme Scalability Mode• ISV support through Cluster Compatibility Mode• High Throughput Computing (HTC)• PGI/Intel compilers• OpenACC Support• GPU-enabled applications
NAMD AMBER
CHARMM GROMACS
NWChem MILC
Matlab
Big Red II – Similar Systems
10
• 11 PFLOPS XE6/XK7 Blue Waters at NCSA
• 20 PFLOPS XK7 - Titan at ORNL
Supporting File Systems – Data Capacitor II
12
• 2 SFA12K40 with 10 84-slot chassis each• 1680 total 3 TB SATA drives – 5,040 TB raw capacity• 16 Object Storage Servers (96 GB RAM) • 2 Metadata Servers (192 GB RAM)
• SFA6620 storage system with 96 600 GB 15K RPM SAS drives & 20 3 TB SATA drives
• 8 Lustre routers – provide access to DCII via a 10GbE• Bandwidth – >20 GB/s via Ethernet; >40 GB/s via InfiniBand
Supporting File Systems – Home Directories
The Home File System uses the DDN SFA12K20E with embeddedGridScaler (GPFS) as NAS storage
Each SFA12K20E has 5 60-slot chassis,140 3TB NL SAS drives for user data and 10 450GB 15K RPM SAS drives formetadata.
Each system has 2 servers for non-NFS client access
Each system has multiple 10 GbEconnections for clients
Native GPFS clients are supported andcan be added if licensed
System 1 -- IUB System 2 -- IUPUI
Condominium computing – on Quarry• Condominium computing - UITS-managed, departmentally
owned research IT systems or system components: • You can purchase nodes that are compatible with IU's Quarry
cluster, have them installed in the very secure IUB Data Center, have them available when you want to use them, and have them managed, backed up, and secured by UITS RT staff.
• You get access to your nodes within seconds of requesting their use.
• When they are not in use by you, they become available to others in the IU community, thereby expanding the computing capability available to the IU community while conserving natural resources and energy.
Questions?
References• PTI – pti.iu.edu• Research Technologies - http://pti.iu.edu/rt• Cray XE6/XK7 - http://www.cray.com/Products/• NVIDIA K20 - http://www.nvidia.com/object/nvidia-kepler.html• Condominium computing - http://kb.iu.edu/data/azof.html• NCSA Blue Waters - http://www.ncsa.illinois.edu/BlueWaters/• ORNL Titan - http://www.olcf.ornl.gov/titan/• DataDirect Networks - http://www.ddn.com/