For the love of the game
By firstname.lastname@example.org on Sep 10, 2009
9 611 262
Now that is one big number...
Hey, I know a larger one you think, for example 9 615 973. But 9 611 262 is actually the highest one now, in the SPECjbb2005 high score list. For a long time I didn't think I was going to see the old world record broken, but here it is.
A couple of years ago, we worked with our friends at SGI and made a series of publications, the highest one of them was 5 180 451 SPECjbb2005 bops (80 945 SPECjbb2005 bops/JVM), which was the new world record. I was saying to myself, "Oh, well, this may actually kill the benchmark". To succeed on SPECjbb2005 you need good software and good hardware. In order to create a new world record in highest total score you need good and large hardware, The result from SGI was run on an Altix machine with 128 sockets, which I am fairly sure is a bit larger than the average closet server.
I was therefore quite happy too see the recent publication from 3Leaf. After too long time without a new world record, the stars were right for a new one. Of course I was happy to see that this world record also was based on JRockit. This also showed me that the "Highest total score" category of the benchmark wasn't dead yet, and the invigorated competition pleased me even more .
During a SPEC workshop we did a presentation about "our love of the game", me from Oracle (at that time BEA), Alan Adamson from IBM and Dave Dagastine from Sun. We all agreed on how much we enjoy the leapfrogging game that had been going on with the benchmark, driving the products forward at a very high rate, to the benefit of all our customers.
And the leapfrogging game continues, now with the most recent publication from SGI with 9 611 262 SPECjbb2005 bops (75 088 SPECjbb2005 bops/JVM). Will this monstrous high score finally kill the competition? Likely not. I know we at the JRockit team will keep improving our JVM, and I also know that the excellent techies at companies like SGI and 3Leaf will keep pushing the envelope. Until the next one.
With love for the game,
SPEC Disclosure Statement
SPEC, SPECjbb reg tm of Standard Performance Evaluation Corporation.
Results as of 09/10/2009 on www.spec.org
Oct 26, 2007 - SGI - SPECjbb2005 bops = 5180451, SPECjbb2005 bops/JVM = 80945
BEA JRockit(R) 5.0 R27.4.0-90 (build R27.4.0-90-89592-1.5.0_12-20070928-1715-linux-ia64, compiled mode)
Altix 4700 (64 Blade/128P DC)
# of Chips 128, # of Cores 256
Jun 5, 2009 - 3Leaf Systems, Inc. - SPECjbb2005 bops = 5534233, SPECjbb2005 bops/JVM = 172945
Oracle JRockit(R) 6 P28.0.0 (build P28.0.0-29-114096-1.6.0_11-20090427-1758-linux-x86_64, compiled mode)
3Leaf Systems Voyager Platform, Quad-Core AMD Opteron(tm) Processor 8384
# of Chips 32, # of Cores 128
Jul 26, 2009 - SGI - SPECjbb2005 bops = 9611262, SPECjbb2005 bops/JVM = 75088
Oracle JRockit(R) (build R27.6.3-40_o-112056-1.5.0_17-20090318-2103-linux-ia64, compiled mode)
Altix 4700 (128 Blade/256 DC), Intel DC Itanium2 Processor 9040 1.6 Ghz
# of Chips 256, # of Cores 512