News Focus
News Focus
icon url

wbmw

04/28/03 6:20 PM

#3290 RE: dougSF30 #3286

Doug, I respectfully disagree with you on a number of points.

1. Opteron outperforms Xeon on many server benchmarks (though not all), but Xeon still seems to have an advantage in many workstation apps due to Opteron's low frequency.

2. Opteron performs closely to Itanium 2 in some tests, but gets vastly outperformed in others. So far, two tests come to mind. In TPC-C, Itanium 2 has a huge lead. In SPECweb99_SSL, they seem about even.

3. As for half the power, I have no desire to get back into that discussion, but I will say that it's arguable. At any rate, it is safe to say that Opteron is lower power.

4. As for half the die size, I don't think that translates into a huge advantage for Opteron, once you take into account yield and bin split problems. I'm willing to bet that Madison has more die per wafer at 1.5GHz than Opteron has at its top speed.

5. As for 1/10th the software porting work, that statement is bunk. Most of the heavy lifting for Itanium 2 has already been done. Using the most advanced development tools for both architectures, you should be able to port an application to either one with equal ease.