Dan, Re: "We have yet to see anything approaching an actual test."
Yes, we have. Having a little trouble recognizing the SAP test? Does this site ring a bell?
sap.com
You previously linked to it several times in a vain attempt to exploit a weakness in the original Itanium, showing it to be on par with older Pentium III Xeon cores. But now the Itanium 2 gains 2.35x in performance over it's older sibling, allowing it to easily outperform competitive entries. Are you no longer willing to acknowledge this as an actual test, just because the Itanium 2 can now perform?
hp.com
Isn't that hypocrisy a little obvious, even for you?
Re: "300 million records is no big deal for any modern database. How well will Itanic do under a real load - such as Anandtech demonstrates when they run their actual site loads against test servers."
Anand's little front end application may have hundreds of forum posters, hundreds of ads flash by, and thousands of page hits in a given day, but that pales in comparison to any backend database load, especially one with 300 million records. (Funny how you mock that size, too, since it is likely hundreds of times larger than anything Anand could possibly have.) Fortunately, though, there are tests that can place the kind of stress that one would find in a high end database, and I linked to one of them in my previous response. Obviously, you hadn't paid close enough attention.
hp.com
Itanium 2 in a 4-way configuration rivals the performance of competitor's 8-way systems, while a 2-way Itanium 2 system rivals the performance of most 4-way entries in this highly stressful database test. Being that TPC is slightly more recognized by the industry than the Anandtech Server Bench, I'm quite sure that Itanium 2 will be positively recognized, despite having missed a review by Anand.
wbmw |
