Canterwood vs. Granite Bay

What is the difference between the 875P and 7205 chipsets when both are running at 200MHz FSB? According to a test by THG, the answer is "not much".

Our tests show that the old Intel 7205 chipset runs stably with a new P4 processor at an FSB clock of 200 MHz. This raises the question of what advantages the new high-end 875 chipset really has over the old 7205. Moreover, in the benchmarks, the two chipsets give about the same results.

Click to expand...

But what about the PAT memory accellerator on the 875?

This makes the PAT technology, highly acclaimed by Intel for its fast memory transfer, seem questionable. In the measurements, we also threw in the Abit IC7-G with the Intel 875 chipset as an example, which performed even more poorly than the overclocked test board with the 7205 chipset.

Click to expand...

BHD

I don't think anyone knows about the performance boost PAT gives - at least, I don't think there were any benchmarks about it. I think that theoretically, it should give a boost of 3-7% at most...

That article at THG comparing the 875 and 7205 chipsets was a bunch of dribble. THG is assuming that all 7205 motherboards will overclock to 200Mhz FSB, I have read reviews where some boards did not make 200Mhz FSB.

Also, THG should be comparing overclocked 7205 boards to overclocked 875 boards. Comparing an overclocked chipset to a non-overclocked chipset is just stupid. Nearly all 875 boards overclock from to 240 – 300 FSB, which is much higher than what 7205 boards can overclock to. If you get one of the new D1 stepping 800FSB P4s, you will want an 875 or 865 board so you can overclock the CPU.

The THG article was a thoughtless and misleading.

from tom and his dreaded bs,he also isnt to clear and percise on what he says,quote "In the measurements, we also threw in the Abit IC7-G with the Intel 875 chipset as an example, which performed even more poorly than the overclocked test board with the 7205 chipset. "
-overclocked granitebay compaired to stock abit??ofcorse the overclocked system will beat the stock,he doesnt mention what was anything clocked at in this statement
stupid thg pos site

Typical THG, all BS

Wired_Monk said:

That article at THG comparing the 875 and 7205 chipsets was a bunch of dribble. THG is assuming that all 7205 motherboards will overclock to 200Mhz FSB, I have read reviews where some boards did not make 200Mhz FSB.

Also, THG should be comparing overclocked 7205 boards to overclocked 875 boards. Comparing an overclocked chipset to a non-overclocked chipset is just stupid. Nearly all 875 boards overclock from to 240 – 300 FSB, which is much higher than what 7205 boards can overclock to. If you get one of the new D1 stepping 800FSB P4s, you will want an 875 or 865 board so you can overclock the CPU.

The THG article was a thoughtless and misleading.

Click to expand...

Welcome to the forums.

I think we read the article differently. I didn't pick up that Tom's was saying all GB boards can run at 200MHz, or that an 865 or 875 would not be a better choice for overclocking a Northwood C proc.

In the last couple of weeks Intel has generated a lot of hype for how the components of the 875 board will be the best of the bins and the PAT memory accellerator. The effect, both implied and stated, is that at equal clock speeds these advantages will give the 875 boards better performance.

This was an interesting, if not conclusive, way to measure those claims. I disagree with you that increasing the board FSB gave the GB a performance advantage. Both boards used the same Corsair XMS running at the same speeds, and at the same voltage, and generating the same bandwidth. Ditto with the processors. Until the 865 boards arrive for testing that's as close to a level playing field as you can get.



BHD

BaldHeadedDork said:

I didn't pick up that Tom's was saying all GB boards can run at 200MHz
BHD

Click to expand...

THG said, "Our tests show that the old Intel 7205 chipset runs stably with a new P4 processor at an FSB clock of 200 MHz." This is a misleading statement suggesting that you can expect boards based on the 7205 chipset to operate at 200Mhz. I wouldn't count on it.

Then THG said, "This raises the question of what advantages the new high-end 875 chipset really has over the old 7205." Again, Tom is misleading people by suggesting that there are no advantages for getting the 875 chipset. Clearly if both boards are overclocked the 875 boards outperform the 7205 boards. Tom didn't bother to overclock both chipsets because he was bent on trashing Intel. He didn't even bother to mention the possibility that 875 boards could operate at higher clock speeds than 7205 boards.

We won't know the benefits of PAT technology until more 865 reviews are done. But early reports seem to suggest that 865 boards that are not overclocking as well as 875 boards.

Any references for those early 865 tests?

BaldHeadedDork said:

Any references for those early 865 tests?

Click to expand...

Somebody from the forums has a MSI 865 board. She's not having an easy time overclocking it. It could be a defective board.

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=189811

If its true that 875 chips have been batch sorted from 865 chips (as stated at Anandtech), it makes sense to believe that 875 boards will overclock higher than 865 boards.

Yeah, I've been following the story of Maxvla's trials with his MSI. I'm sorry to hear about his problems but he's not alone. Ed from the mothersite ran this column today about the problems people are having with the MSI 865 and 875 boards. (Among other Canterwood problems.) Its not the first time this has happened with MSI, either. Last week, I think on the same day Maxvla ordered his board, he wrote:

Of course, with one exception, the "regular" Intel chipset for this platform, Springdale, hasn't shown up yet, and I suspect most people will end up buying that.

Per that one exception, in the past, when that particular company has come out with a mobo long before anyone else, it was less than half-baked. Sorry, but it's "guilty until proven innocent" in my books.

Click to expand...

That company was MSI and I couldn't have said it better myself. MSI has overpromised and underdelivered so many times I will not recommend their boards anymore, or consider their performance in judging a chipset.


I'm almost positive Intel is serious when they say they will cherry pick from the bins for the 875 chipsets. There is too much to lose via a false advertising suit if they don't, and I don't think they want to go through that again.

The question no one has asked is, what is the tolerance that seperates the best Canterwood/Springdale chips from the worst? If the tolerance is big then I agree the 875 would be a better choice for overclocking and just for stability.

But that would be symptomatic of a sloppy and inefficient manufacturing process. If that is the case Intel is also thowing away a bunch of chips not even good enough to be in the Springdale, and opening themselves to warranty liability for those bad chips that get through.

It would also be entirely out of character for Intel to produce chips with a tolerance big enough to create a measurable performance difference. Much of Intel's legendary chipset quality comes from being able to produce millions of chips with the tiniest amount of variation. They are better at it than any other motherboard chipset maker in the world. I don't see that changing for the 865/875 chips.

If the tolerances follow the traditional Intel standard there will only be a small difference between the 865 and 875 in a few benchmarks, and no noticeable difference in application performance. The cynic in me thinks this is Intel's way of adapting the auto industry trick of pushing out the loaded versions of a car when it is launched to capitalize on the immediate demand, and then build less-profitable versions after that demand has fallen off.

But the next month will tell.





BHD