Blog Authors: ELISABETH STAHL 0600029JQ8 is Chief Technical Strategist and Executive IT Specialist, IBM Systems and Technology Group, and has been working in systems performance for over 25 years.
Just Doing My Job
This week HP tells a "Real Story" claiming more performance than IBM on the SPECjAppServer2004 benchmark. And if they did their job, they would have told the whole story.
You see, for this benchmark, it is important to look at all the data. That means the database server hardware and software as well as the application hardware and software. And if you look at the database server you would see that HP needed over 3x the number of database cores that IBM needed to get only 1.5x the performance.(1)
To do the benchmark analysis job well, you need to examine all of the data available. Not just the parts that make your story more real.
This week President Obama spoke with the pilot who skillfully landed the Airbus 320 in the Hudson River. The pilot told him, "Me and my crew, we were just doing our job."
"And it made me think, if everybody did their job -- whatever that job was -- as well as that pilot did his job, we'd be in pretty good shape," Obama said.
(1)HP SPECjAppServer2004 result of 21,602 JOPS on an HP Integrity BL870c server blade dual-core Intel® Itanium® 9100series 1.6GHz, 13 blades/52 chips/104 cores. Each blade: 8 cores/4 chips/ 2 cores per chip running Oracle WebLogic Server Standard Edition Release 10.3 HP–UX 11iv3 September 2008 . The database server was a Superdome configured with dual-core Intel Itanium 1.6 GHz processor chips (128-core). vs. IBM SPECjAppServer2004 result of 14,004 JOPS using WebSphere 6.1 Application Server with EJB3 Feature Pack on IBM System pSeries Blade Center and DB2 9.5 on a configuration with 26 application servers, each an IBM BladeCenter JS22 Express (2 4.0 GHz POWER6 chips, 4 cores, 2 cores/chip, 10 GB memory) running WebSphere 6.1 Application Server with EJB3 Feature Pack, IBM J9 VM on IBM AIX 5L V5.3 and the IBM System p5-595 database server(40 cores, 20 2.1 GHz POWER5+ chips, 2 cores/chip (SMT on), 192 GB memory) running IBM DB2 9.5 on IBM AIX 5L v5.3. Source: http://www.spec.org. Results current as of 1/23/09.
SPEC, SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPECompM, SPECompL, SPECsfs, SPECpower, SPEC MPI and SPECpower_ssj are trademarks of the Standard Performance Evaluation Corporation (SPEC).
The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.
technorati tags: IBM, systems, performance, HP, benchmark,SPEC,SPECjAppServer2004,DB2,JS22,p5-595,WebSphere,benchmarking