differences between 965 and 975?

[Torin]

975 is the enthusiast chipset and is the only one to support extreme edition processors, at least the boards made by Intel. If the intel made 965 boards won't support extreme edition why would the 965 boards made by other companies?

Why would you think the 965 boards won't support EE chips?

They do, look here: http://www.anandtech.com/cpuchipsets...oc.aspx?i=2771

Chipset supports Conroe, why would Conroe EE be any different?

[pik-ard v1.1]

lol that rabbit is gettin to you man. haha

Don't the current USB and Firewire Drives have Extermal power source ?... same for external SATA..

Can we start saying good bye to USB/Firewire? who needs 400MB/s when you can have up to 3GB a second

yea... i think it may be... i dunno why i messed up so much thinking about that one little sata port

hmm... and i just got a usb thumb drive. i wonder if sata will get to the point where there will be lightning fast sata thumb drives.. would be cool, though the design isn't as compact as usb ports. though of course they could rearange it just like there are more compact usb connectors

[FischOderAal]

My thoughts exactly...it may work, but how well is anybody's guess. Sort of like the old Xeon MP, running 4 CPUs on 1 lousy 100MHz bus...

Of course, running two Clovertown on a Bensley board would result in a similar sort of bottleneck...albeit on a factor of two (one for each socket/bus). Each half of a Clovertown would roughly be equal to a Merom running on Napa (333MHz/2). But, as you say, the bottleneck would be more acute on the desktop side due to the 266MHz bus vs the 333MHz bus for Bensley.

What may work to alleviate this bottleneck would be something like a DDR2-800 situation with a 1:2 bus:memory configuration. Broadwater is bringing DDR2-800 to the table, so really all that would be needed is validation at 400 MHz, which considering many chipsets are running there already, shouldn't be entirely out of the realm of possibility.

Here's hoping, in any event...

yepp, Intel's Quadcore will lack performance due to the missing IMC (my opinion). that's why Intel introduced Dual FSB in Server boards (afaik).

but many 975X Boards have native DDR2 800 support, too. for example the P5WD2-E, P5WDG2-WS, P5W-DH... just to mention the ones of Asus.

well, but I still seriously doubt that 965 will beat the 975... I think Intel would rather bind a stone at the limbs of 965 than see the 975 being beaten by the 965...
that would be like a 7900GT beats a 7900GTX...

[Torin]

well, but I still seriously doubt that 965 will beat the 975... I think Intel would rather bind a stone at the limbs of 965 than see the 975 being beaten by the 965...
that would be like a 7900GT beats a 7900GTX...

The 975X is a Crossfire chipset, what is so hard to believe about a more barebones chipset (P965) beating it or at least equalling it in performance? You can't compare chipsets to GPU models... the GTX has faster and more memory, that's why it performs better, no similiar parallel can be made between 975X and P965.

[Gautam]

/sigh.... I don't understand why i have to keep saying this. 965 WILL outperform 975. I've seen the numbers, I've talked to the design engineers, I know what I am saying. Intel is pulling out all the stops on this one. It has little to do with only DDR2-800 support.... it has to do with the way the MCH works in 965 as opposed to 975. It's a newer, "better" chipset.... no way around it.

So, what do you have in store for us Crossfire users?

[FischOderAal]

/sigh.... I don't understand why i have to keep saying this. 965 WILL outperform 975. I've seen the numbers, I've talked to the design engineers, I know what I am saying. Intel is pulling out all the stops on this one. It has little to do with only DDR2-800 support.... it has to do with the way the MCH works in 965 as opposed to 975. It's a newer, "better" chipset.... no way around it.

the same way you "knew" Kentsfield wouldn't boot on 975? sry for that one...

[FischOderAal]

Wow... you're an idiot. If you look back I always said that I wasn't 100% sure on that one and that I would verify it. Guess what... I'm 100% sure on this one.

no need to call someone an idiot if that's your niveau...

if you know you're right, fine. I only say that for marketing reasons for example this would be bull and we've seen often enough that some products had integrated breaks, just to make them slower...

[Cooper]

Like I care.....

With that attitude you just might go outside

*NO ABUSING*

[Cooper]

WTF? Yea, cause what he said was SOOOOOOO polite. Sorry if I don't hide my contempt for people in subtext...

There`s no ofence in his statement. And please do try to behave yourself. Any misunderstandings take to PM.

[mine]

/sigh.... I don't understand why i have to keep saying this. 965 WILL outperform 975. I've seen the numbers, I've talked to the design engineers, I know what I am saying. Intel is pulling out all the stops on this one. It has little to do with only DDR2-800 support.... it has to do with the way the MCH works in 965 as opposed to 975. It's a newer, "better" chipset.... no way around it.

I am pretty sure that you know what you are saying this time ...
but would you do us a favor and
comment a little bit on the advantages of the 965 MCH ? " I know " isn`t very describing .
As far as I remember you refer to the "Intelfast memory Access" feature
and the DDR2-800 support ..
now ....
I see the first anand review of a 965 based mobu/ Conroe combi and guess what , I have some problems to see the big step ....

Thanks in advance for a little lesson in chipdesign...

[Cooper]

So you mean the only advamtage of 975 over 965 is CrossFire support ?
Well, chipset been officialy presented - perhaps you might give us some comparative numbers (or at least %) ?

[Cooper]

That`s a very nice boost. Conroe and specially Kentsfield would definatly benefit from that.
Anybody knows when will 965 based mobos will hit the market ?
Something tells me that`s gonna happen very soon and we`ll see a lot of comparative tests with 975 ones

[Plaicd]

Intel P965 is heavy on the juice.
http://www.theinquirer.net/?article=32230

North Bridge 975X 945P P965
South Bridge ICH7 ICH7 ICH8
Bus Speed 1066MHz 1066MHz 1066MHz
Memory 667 x 2 667 x 2 ICH8
NB TDP 13.5W 15.2W 19W
SB TDP 3.3W 3.3W 4.1W
Total TDP 16.8W 18.5W 23.1W
----------------------------------------

No idea how this might effect its overclock but maybe something to consider.

[mine]

20% increase in memory performance over 975. That's a significant increase, more than just DDR2-800 would provide.

I will remind you , when first reviews pop up

[FischOderAal]

That`s a very nice boost. Conroe and specially Kentsfield would definatly benefit from that.
Anybody knows when will 965 based mobos will hit the market ?
Something tells me that`s gonna happen very soon and we`ll see a lot of comparative tests with 975 ones

I don't think Conroe and Kentsfield would benefit from that. because Ram-Performance is good enough, the FSB is the Bottleneck. and unless we have Dual-FSB or IMC this Problem won't be fixed.

just my 0,02$ ^^

ich will remind you , when first reviews pop up

@mine: leicht verwirrt ^^ denglisch

Intel P965 is heavy on the juice.
http://www.theinquirer.net/?article=32230

North Bridge 975X 945P P965
South Bridge ICH7 ICH7 ICH8
Bus Speed 1066MHz 1066MHz 1066MHz
Memory 667 x 2 667 x 2 ICH8
NB TDP 13.5W 15.2W 19W
SB TDP 3.3W 3.3W 4.1W
Total TDP 16.8W 18.5W 23.1W
----------------------------------------

No idea how this might effect its overclock but maybe something to consider.

the 965 is 0,09nm (975 is 0,13). might be Prescott Effect ^^ smaller structures, but more leakes and because of that more heat!
I think it's funny that the 965 actually has increased the tdp, while Intel pronounces that the Broadwater will make your system more quiet

[FischOderAal]

ah, Poncho. what do you say about this article?
they used the Gigabyte P965-DQ6 with i965.
http://www.tweaktown.com/articles/91..._pi/index.html

and compare it to "xxmartins" Benchmarks on Bad Axe... (they used DDR1066, xxmartin used DDR800 - okay, xxmartin had 4-4-4 timings on them)
http://www.forumdeluxx.de/forum/showthread.php?t=237050

what do we see?
SuperPi? exactly the same... 21 secs (and I thought Latency is very important for this one)
Sciencemark 2.0:
Molecular
P965 64,79 sec and Bad Axe 65,045 sec (which is damn near and should be due to benchmark-inaccuracy)
Primordia
P965 267 sec and Bad Axe 271,59 sec (3 secs difference... not a "significant amount" and might be due to inaccuracy as well)

to bad they didn't test Sandra Bandwith.

I think we will see more Benchies which will tell the same...

[JoeBar]

Let's way and see. Soon enough we're going to clear this out.