Problems with Xenon article
Just out of curiosity, why is it that in the new Xenon bench tests they sometimes compare 4 Opteron 844's, 2 Opteron 244's, and other times just regular Athlon processors? Some consistency would be nice, unless we are trying to paint a specific picture...
And there you've found the point of the article. What they really should do is have a Xeon video of how it's performance relates in doing multiple tasks. I use many different Xeon based machines as well as the latest p4's here at work, and for doing db work and coding, the Xeon has visibly faster results.
I would still buy a p4 3.2 if i had the chance, because it is running at the higher bus and is cheaper, but once the Xeon moves to 800, i think there will be a shift. This isn't canabilization or a lack of competition with AMD as the article seems to suggest. Intel seems to be waiting on the 800 Mhz FSB Xeon. They just upped the cache (hence the reason for the article), but the still are limiting the bus. I wonder how much heat a dual 3.2 Xeon system would create (and maybe that's an issue). *shrugs*
But i do think the article was a little off base. I've read a few other Xeon articles that actually kept the systems consistent and found them to be much more useful.
If the article's title was supposed to be an indication of the content, i say they should have used but 2 systems - dual system with old xeons, dual system with new, higher cache xeon. The way it is now is just confusing...
I'm just your average habitual smiler =D
They really needed to include webserver benchmarks, they would benefit more from the added cache than stupid stuff like encoding and games(which people won't be doing most of the time with an expensive dual setup), which should be left out of Xeon reviews, if you want to review xeons with gaming do it in a seperate review
I agree. There should have been a lot more consistency. I mean, that sure was "fair" of THG to compare a 4 Opteron system to a dual Xeon system. Plus, I think there should have been some webserver benchmarks, and some <i>actual</i> workstation benchmarks. It seems that THG uses the same old benchmarks, whether it's a server CPU in review, or a desktop CPU.
- - -
"... In the semiconductor industry, it's good to be paranoid ..." - [Andy Grove]</font color=green>
Actually its technically 666.666 recouring, so its averaged to 667. And sounds less demonic.
IIRC the celleron at that speed was also a 667Mhz part :wink:
I remember seeing an intel roadmap with bus speeds ages ago.
starting at 400, then 533 then 667 then finally 800 with DDR2. Obviously they skipped a speed grade with the advent of high speed DDR1 and the delay of DDR2.
ECC PC3200 is gonna be hard to make methinks.
Mr no integrity coward.</b>