Contribution Details |
| |
| Name: |
 |
Gordon: Design, Performance, & Experiences Deploying & Supporting a Data Intensive Supercomputer |
|
| |
| Time: |
 |
Wednesday, June 20, 2012
11:30 AM - 12:00 PM |
|
| |
| Room: |
|
Hall 4
CCH - Congress Center Hamburg |
|
| |
| Speakers: |
|
Shawn Strande, San Diego Supercomputer Center |
|
| |
| Abstract: |
|
The Gordon data intensive supercomputer entered service in early 2012 as a computing system in the NSF Extreme Science and Engineering Discovery Environment (XSEDE) program. Gordon has several innovative features that make it suitable for data intensive computing including: 1,024, dual socket, 16-core, 64GB compute nodes based on Intel's Sandy Bridge processor; 64 I/O nodes with an aggregate of 300 TB of high performance Intel 710 flash (SSD); large, virtual SMP "supernodes" of up to 2 TB DRAM based on ScaleMP's vSMP Foundation; a dual-rail, QDR InfiniBand, 3D torus network based on Mellanox hardware and open source software; and a 100 GB/s Lustre-based parallel file system, with over 4 PB of high performance disk. In this talk we present: the motivation for and design of Gordon; performance on a range of micro and application benchmarks; some promising early results in application-aware energy savings techniques; and highlight the challenges and risk mitigation strategies in deploying a data intensive supercomputer like Gordon which embodies significant innovative technologies. Finally, we present our experiences thus far in supporting users and managing a system like Gordon. |
|
|
|
|
|
