> I respectfully disagree (IRD). I am building a > cluster to support high-performance computing, > specifically meteorological models. The current > version of the code runs on UltraSparc-III non-copper > based system. The Sun sales rep loaned us an x4200 > and I borrowed times on a SunFire V490 both > configured with 4 cores and 4gb or memory. The > wall-clock run time for the x4200 was 15 hours and > the v490 was 9 hours. Granted the x4200 was cheaper, > but you also get cheap hardware. The x4200 has too > many sharp edges and reminds me of a beige box where > the disk drives are held in place with foam blocks.
Please note that I was specifically referring to the desktop/workstation space, and that I stand by my previous statement that Sun has nothing of interest or performance in that segment. Even a 32-bit Pentium 4 system with Hyper Threading will blow the doors off of any UltraSPARC workstation currently available for sale by Sun or any alternative SPARC hardware manufacturer. Also please note that X4200 and the V490 are both server class hardware, which is a completely different arena. Additionally, a V490 is equipped with the latest generation UltraSPARC IV+, dual core processors and is the Sun's current number-crunching UltraSPARC processor, at least until Niagara II and Rock are out. So in comparing X4200 and V490 with the desktop offerings, you're really comparing oranges with apples. If on the other hand you're trying to communicate that a V490 is faster than an X4200, that information is really meaningless without giving more details. For example, you could have been comparing the results of a run on a dual core UltraSPARC IV+ versus a single core 200-series Opteron processor. Not a fair comparison. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
