I’ve been working with the P4P800 Springdale motherboard the last couple days after working with the Abit IC7.
I wish I could jump on the Springdale bandwagon like just about everyone else and tell you that there’s no reason to even consider a Canterwood motherboard.
I can’t. I just don’t feel comfortable doing so given my experiences with the two.
To compare the two, I chose a very simple testing methodology. Others have already run a million benchmarks at spec which indicate that the two are basically equal. Why reinvent the wheel? This website isn’t called spec.com; it’s called overclockers.com. I’m interested in overclocking
So what I do is establish speed and stability. First, I run FSB/memory at 3:2 to keep memory from becoming the bottleneck, and crank it up as far as I can go while keeping stability, primarily under Prime95. That way, you’re testing the chipset, and only the chipset.
For the Springdale, I turned on “PAT.” For the Canterwood, I turned on no performance tweaks simply because they led to instability at any decent speed. (The memory benchmarks came out about identical for both under these circumstances.)
Canterwood 295, Springdale 289.
Next, I did the same using an FSB/memory ration of 5:4.
Canterwood 280, Springdale 276.
It’s close. Factor in cost, and it may be close enough for you to buy a Springdale rather than a Canterwood, and that’s fine.
But from what I’ve seen, I can’t pretend that the Springdale chipset is quite as good, much less better than the Canterwood, which is the impression that’s often being left nowadays.
A Few Words On Springdale Heat
Unlike the IC-7, the P4P800 comes with only a passive heatsink. This thing gets hot at default speed, and at 280+, you’ll literally get burned by mine with any prolonged contact.
However, placing a fan on it didn’t seem to help maximum performance very much (though it cooled down the heatsink considerably); I could only get an extra MHz FSB from it. Then again, when you have a motherboard out in the open in a cool room with a water-cooled CPU, that represents more ideal conditions than can be found in the average case.
As a precaution, if nothing else, a simple fan remains a good idea.
On Any Given Sunday . . .
. . . one NFL (American football) team can beat another. Springdale vs. Canterwood represents two closely matched NFL teams competing against one another.
Of course, this is the case of one man testing one Canterwood and one Springdale mobo. There are those who have found under different circumstances that Springdales do a bit better, and that doesn’t surprise me at all.
If a hundred of you did exactly the same thing I did, I bet some of you would find the Springdale better, and some would find even a bigger Canterwood gap.
This is a sure sign it’s close between the two. So long as there is little difference between the two, even the “inferior” product will win sometimes due to normal variation.
This is why comparative reviews of closely matched products is statistical nonsense, and why we don’t like doing them. We find them inherently misleading (even when we do them). The truth lies in a murky statistical cloud, and the decisive factor for you may not be “Which type of board is better?” but rather “Is this particular motherboard in my hands better than that particular board on the table for the equipment I’m going to stick into it?”
The only way to find that out is to try it.
Is Intel Full of It or Not?
Intel is not happy with what’s going on here. What concerns me is that people are automatically assuming, “Intel must be lying about these chipsets. It’s all part of their conspiracy to rip us off, but we’re too smart for that.”
In short, believing what you want to believe. I get very uncomfortable when conclusions match wishful thinking.
I’m nowhere near as confident about that.
It really boils down to whether or not Intel is effectively binning. Do they really test everything, and only the cream of the crop ends up being Canterwoods? Or (more likely) do they just test enough to get their quota of Canterwoods, and everything else that works becomes Springdales, and of the Springdales, how many of these would qualify as Canterwoods?
We don’t know.
In short, are you actually paying more for a higher quality chipset, or just for the validation that it’s a quality chipset? And just how relevant is the validation to an overclocking situation, anyway?
All we can say for sure is that a Canterwood passed some sort of test. A Springdale either failed the test or (more likely) wasn’t tested.
Does it matter? I don’t know. I’m neither confident enough to say that those who think Springdales = Canterwoods are either right or wrong.
When we will know if there’s a real difference or not? Maybe in a few months. That’s not going to help the guinea pigs now.
The Paragraph That Fuels My Fear
Others have also looked at the P4P800. Here’s what one reviewer had to say:
“In terms of overall board stability, the P4P800 was rock solid up to 300 MHz PSB but completely failed above 302 MHz bus speed. Kyle from [H]ardOCP had emailed me about stability problems above 250 MHz he experienced with his board unless active cooling was added to the MCH, we did not see anything like this, moreover, the MCH heatsink never got anything but lukewarm to the touch. Whether there is a story behind this or it is just coincidence is completely beyond our knowledge and we won’t comment further.”
That paragraph bothers me, and it should bother you. One person can’t get above 250MHz without additional cooling. Another hits 300MHz with a lukewarm heatsink. I get 290MHz, but the only place you’d call my heatsink lukewarm would be in the neighborhood of a blast furnace.
This tells me there’s a lot of variation among these Springdales, which means there’s a level of risk associated with them. Not a big deal running stock, and probably not one running a modest overclock, but a much bigger deal when you are pushing the envelope.
Is the Savings Worth It?
It would be one thing if the cost difference between the Springdale and Canterwood was huge, but it isn’t.
The cost difference between the P4P800 and the IC7 is about $15. That’s not a lot. Even if you compare the IS7 to the IC7, the cost is about $35.
I know full well that nothing I say short of presenting concrete proof that Springdale chipsets emit nerve gas when run above 250MHz will dissuade most people from saving $15 or $35.
They may well end up being right (or maybe most of them). Then again, they may end up being wrong.
This article isn’t for the bold (or foolish, only time will tell which it is). It’s for the cautious, for those whom a few extra dollars isn’t critical, those who hesitate to jump on the Springdale bandwagon.
To those people I say, “It’s OK to spend the extra money; it’s not foolish. It may pan out, it may not, but there’s some reason to have doubts about the Springdales being as good or better than the Canterwoods.
Especially for those who want to push the envelope.