Using SPARC and Solaris for HPC: More of this, please!

Ken Edgecombe – Executive Director of HPCVL spoke today at the HPC Consortium Meeting in Austin about experiences with SPARC and HPC at his facility.

HPCVL has a massive amount of Sun gear, the newest of which includes a cluster of eight Sun SPARC Enterprise M9000 nodes, our largest SMP systems. Each node has 64 quad-core, dual-threaded SPARC64 processors and includes 2TB of RAM. With a total of 512 threads per node, the cluster has a peak performance of 20.5 TFLOPs. As you'd expect, these systems offer excellent performance for problems with large memory footprints or for those requiring extremely high bandwidths and low latencies between communicating processors.

In addition to their M9000 cluster, HPCVL has another new resource that consists of 78 Sun SPARC Enterprise T5140 (Maramba) nodes, each with two eight-core Niagara2+ processors (a.k.a. UltraSPARC T2plus). With eight threads per core, these systems make almost 10,000 hardware threads available to users at HPCVL.

Ken described some of the challenges of deploying the T5140 nodes in his HPC environment. The biggest issue is that researchers invariably first try running a serial job on these systems and then report they are very disappointed with the resulting performance. No surprise since these systems run at less that 1.5 GHz as compared to competing processors that run at over twice that rate. As Ken emphasized several times, the key educational issue is to re-orient users to thinking less about single-threaded performance and more about "getting more work done." In other words, throughput computing. For jobs that can scale to take advantage of more threads, excellent overall performance can be achieved my consuming more (slower) threads to complete the job in a competitive time. This works if one can either extract more parallelism from a single application, or run multiple instances of applications to make efficient use of the threads within these CMT systems. With 256 threads per node, there is a lot of parallelism available for getting work done.

As he closed, Ken reminded attendees of the 2009 High Performance Computing Symposium which will be held June 14-17 in Kingston, Ontario at HPCVL.

Comments:

It seems counterintuitive that researchers at an HPC facility would not use parallel code from the get-go on these machines. If HPC researchers are still in the serial coding mindset; it is going to take a while for parallelism to filter out to the software world generally.

Posted by Scott on November 17, 2008 at 08:25 AM EST #

On the one hand it does, but we are also seeing new people entering HPC who may not have the expertise to deal with parallelization. It seems that more people are now interested
in extracting more performance from their codes...and for
multicore processors that often means parallelization where
before serial codes would just become faster as clock speeds
increased. That free ride is over.

Josh

Posted by Josh Simons on November 18, 2008 at 01:21 PM EST #

Post a Comment:
  • HTML Syntax: NOT allowed
About

Josh Simons

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
   
       
Today