Running PIV faster may not mean they perform better.– Tim Cole
I’m seeing some odd results from this Northwood of mine.
It’s running at 2.8GHz just fine, but what it’s not doing is producing noticeably
better results than my older 2GHz willy.
I run SETI 24/7 on everything I have, so I have a good feel for it. I compared the Team Lamb Chop bench
score from my old 2GHz chip to that just run on this box at 2.8GHz, and the
difference was only 23 minutes (Ed.note: a little more than 10% improvement). This is not at all what I expected from a 40%
MHz jump. I get the feeling something is not quite right. All of the chipset
drivers are installed, and the board seems to be tweaked as much as
What really raised a flag is I got this 1.8A part running on an
i845 board, and while cruising at 2.4GHz it’s outperforming the 2.8 box in
some areas. Can’t put my finger on it yet, but something’s wrong. I was just thinking that maybe it looks like this 2.2 is
running fine at 2.8, but perhaps it’s actually getting hot enough to trigger
the throttling diode under load and then it’s not running as fast as stock,
or something of that nature.
Ed. note: I have been wondering about this ever since I saw the Tom’s Hardware benchmarking
with an overclocked Northwood. In that case, though, temperature was not an obvious factor.
There’s a few other possibilities. Many of these PIV boards allow for asynchronous settings of FSB and memory. Asynchronous modes in the past have led to performance hits, that may be a factor here.
In the case of i850 boards, if RDRAM is being slowed down, that could lead to a possible bottleneck.
In the case of at least video-intense benchmarks, it may just be a matter of a video card bottleneck.
Or it may well be some sort of throttling is going on.
If you are one of the Northwood pioneers, I’d like to hear from you to see whether or not you’re running into the same thing in your favorite benchmarks when you crank it up. Please include items like the mobo you’re using, the type of RAM, the FSB and memory speeds.