Earlier this week I had some invitations to mail that should have been easy. You use the automated machine, weigh them, check off that they are square and rigid, and you're on your way. Unfortunately the scale kept changing the weight, the scale varied from other scales, and the postal employees then got involved and argued with both the automated machine and themselves about whether each invitation should be 62 cents or 79 cents or one dollar. Was it just a square letter, a large envelope with special needs, or even a package?
My next trip was just as exciting. I only hope my tax forms in multiple sized envelopes to multiple agencies make it. Next year, it's electronic or bust.
And that's how I felt about Sun's latest Intel announcement benchmarks. Too many "world records" once again where you can't even tell whether they are square or rigid. Let's look at just a few.
Sun's executive vice president, Systems Group, states that today's new Sun systems "will give the data center an extreme makeover." Putting automated machines that incorrectly calculate your postage in post office lobbies and taking away valuable service - that's an extreme makeover too that just hasn't been working. Coming soon on ABC - Extreme Makeover: Postal Edition, Tuesdays at 8, Extreme Makeover: Data Center Edition, Wednesdays at 9.
(1)Sun Fire X4275 TPC-H result (2 processor chips/8 cores/16 threads) of 23,365.3 QphH@1000GB, $2.41/QphH@1000GB, available 5/1/09 vs. Fujitsu 1TB TPC-H result of 1,018,321.9 QphH@1000GB ($1.18 per QphH@1000GB , avail. 08/01/2008) obtained on a cluster of 48 Primergy RX300 S4 servers for a total of 96 Quad-Core Intel Xeon X5460 3.16 GHz processors (96 processor chips, 384 cores, 384 threads). (2)Sun Fire X4270 SAP SD 2-tier result (2 processors / 8 cores / 16 threads), Solaris 10, Oracle 10g, SAP ERP 6.0 (Unicode) results: 3,700 SAP SD Benchmark users, 20,300 SAPS. SAP certification number 2009005 vs. IBM System x3650 M2 result of 5,100 SAP SD Benchmark users, 1.98 seconds average dialog response time, 25,530 SAPS, measured throughput of 1,532,000 dialog steps per hour (or 510,670 fully processed line items per hour), and an average CPU utilization of 99 percent for the central server. Configuration: two Intel Xeon X5570 processors, 2.93GHz with 256KB L2 cache per core and 8MB L3 cache per processor (2 processors/8 cores/16 threads), 48GB of memory, 64-bit DB2 9.5, Microsoft Windows Server 2003 Enterprise x64 Edition, and SAP ERP 6.0 (certification number 2008079). (3)Sun Blade X6270 server module (2 processor chips/8 cores/16 threads) SPECjbb2005 result of 503,675 SPECjbb2005 bops, 503675 SPECjbb2005 bops/JVM vs. IBM BladeCenter HS22 result of 604,417 SPECjbb2005 business operations per second (SPECjbb2005 bops) and 151,104 SPECjbb2005 bops/JVM. This result was with two Xeon X5570 (2.93GHz, 2 chips, 8 cores, 4 cores/chip) with 12GB memory.Sources: www.tpc.org, www.spec.org, www.sap.com/solutions/benchmark. Results as of 4/14/09.
TPC-C ,TPC-H, and TPC-E are trademarks of the Transaction Performance Processing Council (TPPC).
SPEC, SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPECompM, SPECompL, SPECsfs, SPECpower, SPEC MPI and SPECpower_ssj are trademarks of the Standard Performance Evaluation Corporation (SPEC).
SAP, mySAP and other SAP product and service names mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries all over the world.
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, TPC-H, benchmark,SAP,SPECjbb2005,Nehalem,Sun,X4270,X4270,X6270,Intel,benchmarking