differences between 965 and 975?

[pik-ard v1.1]

what do we know so far? i cant seem to find much on this.

975 is a bigger number, so i'd expect it to be the top of the line, but will 965 be good enough to get the same ballpark 3.5-3.9GHz on air out of the E6600? or will it overclock slightly less than the 975?

i've never used intel (well, not since 266MHz days), so i dont know how the upper-lower range chipsets compare. :slap:

[vapb400]

The differences are as follows:
965 is newer so it has had some improvements made. It looks like it will be the winner between the two WHEN NOT USING CROSSFIRE/SLI. (as it apparently has its fair share memory controller tweaks) It has the ICH8 southbridge as well as native support for DDR2-800. It however does not have an optimal pci-express lane setup. It is 16x/4x instead of the ideal 8x/8x or 16x/16x thus it is rumored that ATI will not be making Crossfire available for it. Not that a person with a little creativity couldn't circumvent that restriction.

We don't know anything about the overclockability but by nature I would say the 965x SHOULD overclock better, however reality may be alot different. The 965 is not being marketed as the enthusiast chipset, so therefore many motherboard manufactures will not market 965x motherboards as enthusiast motherboards. If that is the case then we will probably see 975X overclocking higher.

I would not put 3.5-3.6ish out of range, the Asus offerings are looking very nice and I'm sure alot of companies will be putting out more than decent clocking motherboards.

[pik-ard v1.1]

thanks. so if all it's technically missing is sli/xfire, then the heck with that because i'm not rich enough for that anyway.

[kyleslater]

The 975 has the 8Phase power (still droops) that the 965 does not have so I will probably be looking at the 975 myself. Also... The Bad Axe 2 is supposed to support Kentsfield so I may get a revision of the 975X that supports that... (Asus of course).

[vapb400]

As far as I know Quad Core will only be supported on 965 for desktop. I've yet to see any validation being done on 975, doesn't mean that it isn't being done though. I'll try and boot a kentsfield up on 975 silicon next week.... but everybody that I have talked to says 965 or newer so I wouldn't hold my breath.

Really though... I would NOT base my decisions on quad core. Until there is a change in the FSB then quad core is a waste IMHO. You get away with it on the Server side since they will have dual independent FSB's and FBDIMMs..... desktop there isn't the bandwidth to take advantage of it. Again... IMHO.
That is quite interesting coming from you. The inq had a pic of a bad axe + what the claimed was kentsfield, but take that with a pinch of salt. FCG has also said that bad axe would take it with a simple bios update

we'll see.

[kyleslater]

Bad Axe 2. Not the original Bad Axe.

[FischOderAal]

well... the 975 still is the High-End Chipset, so I don't think that the 965 will beat the 975 by far ;)

but, well, we don't have any benchies or O/C yet :D let's wait. I will buy a P5W-DH. the P5B Dlx will be released together with the P5W-DH. so I think there will be some comparisons soon (to see how much the Memory Tweaks will increase Performance).

[boostedevo]

Kentsfield will be "officially" released on the 975 chipset from what I've read, so I am not surprised it fired on your 975 board. I am surprised to hear someone has one already though :)


The 975 has the 8Phase power (still droops) that the 965 does not have so I will probably be looking at the 975 myself.ASUS P5B 965 board is 8-phase.

[pik-ard v1.1]

so the boards to get will be the ultra/pro/deluxe editions of the 965 offerings?

[boostedevo]

Tough call, but I wouldn't have a problem buying one to check out since I am running a single card, but 975 is the clear answer for CF.

@FLG: thanks for the info. Is that a new step 975 in that server board? Isn't NDA up on the 965 chipsets today? If so, where are the boards?? My 955 is getting tired :)

[SMa]

Don't you guys think ATI will offer a better alternative?

[FischOderAal]

Everything that I have seen suggests otherwise. Believe me, the 965 chipset will outperform 975 by a noticable amount.....

hmm

what about the NDA? still up? when do you think we will see some benches? I really can't believe that Intel's High End Chipset shall be outperformed by a smaller one... that would be nonsense

[Torin]

hmm

what about the NDA? still up? when do you think we will see some benches? I really can't believe that Intel's High End Chipset shall be outperformed by a smaller one... that would be nonsense
High end in terms of features, not necessarily performance. The 965 won't do CF like the 975X. Extra features are often a burden when it comes to performance, not to mention it is a newer chipset, regardless of how many features it has.

[FischOderAal]

Everything that I have seen suggests otherwise. Believe me, the 965 chipset will outperform 975 by a noticable amount.....


Motherboard manufacturers are particularly interested in releasing 945 based motherboards with Conroe support simply because the performance is apparently close to 965 platforms, but at a much lower cost.
http://www.anandtech.com/tradeshows/showdoc.aspx?i=2770&p=2

who's right? :confused:

[TL1000S]

who's right? :confused:

I would think someone with access to pre-production (almost a year in front) CPU's from the manufacterer would have the "edge" in this matter...
But.. I could be wrong.. just like "insider" FLG_Poncho was doubting a "Kentsfield" would boot on a 975x board.. :)

[XS Janus]

Don't you guys think ATI will offer a better alternative?

If by the term "better" you mean more adventure prone, than sure:p:
________
Anal Webcams (http://www.girlcamfriend.com/webcam/anal-sex/)

[JoeBar]

The 975 has the 8Phase power (still droops) that the 965 does not have so I will probably be looking at the 975 myself. Also... The Bad Axe 2 is supposed to support Kentsfield so I may get a revision of the 975X that supports that... (Asus of course).
That's not the case with Asus P5B Deluxe (http://www.asus.com/products4.aspx?l1=3&l2=11&l3=307&model=1179&modelmenu=1) 965 based board... ;)

[pik-ard v1.1]

hmm, just lookin over the P5B Deluxe...

back pannel (duh?):
http://www.asus.com/999/images/products/1179/p5b-d_back.jpg

what's that port under the firewire? and i like how the firewire is red. caus ya know... fire = red.... or something.

odd that it's only 4 USB ports.. my ultra-d has 6 on the back. but in the description it does say it supports 4 more (2 front + 2 in a back pannel card doohicky?)

[n91htmare]

hmm, just lookin over the P5B Deluxe...

back pannel (duh?):
http://www.asus.com/999/images/products/1179/p5b-d_back.jpg

what's that port under the firewire? and i like how the firewire is red. caus ya know... fire = red.... or something.

odd that it's only 4 USB ports.. my ultra-d has 6 on the back. but in the description it does say it supports 4 more (2 front + 2 in a back pannel card doohicky?)

Looks like an external SATA port. Maybe for future External devices that support SATA. SATA is Hot plugable, so maybe external hard drives that are SATA.

[pik-ard v1.1]

heh. thought it looked familiar... just couldn't place it caus i'm not used to seeing them on the back.


Back Panel I/O Ports

1 x PS/2 Keyboard
1 x PS/2 Mouse
1 x Serial port
1 x S/PDIF Out (Coaxial + Optical)
1 x External SATA
1 x IEEE1394a
2 x RJ45 port
4 x USB 2.0/1.1
1 x WiFi-AP Solo™ antenna jack
8-Channel Audio I/O

guess i should checked that list...

EDIT: hmm... but dont sata drives need seperate power connectors? though maybe there's some new sata thing that provides power... or maybe it splits and connects to both. or maybe we'll just wait and see.

[Torin]

An external SATA drive would probably have it's own wall-plug power adapter similiar to how external IDE drives get their power.

[JoeBar]

Yes it has its own power supply.

[n91htmare]

heh. thought it looked familiar... just couldn't place it caus i'm not used to seeing them on the back.



guess i should checked that list...

EDIT: hmm... but dont sata drives need seperate power connectors? though maybe there's some new sata thing that provides power... or maybe it splits and connects to both. or maybe we'll just wait and see.


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

[lutjens]

Really though... I would NOT base my decisions on quad core. Until there is a change in the FSB then quad core is a waste IMHO. You get away with it on the Server side since they will have dual independent FSB's and FBDIMMs..... desktop there isn't the bandwidth to take advantage of it. Again... IMHO.


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...:rolleyes:

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...;)

[gone_fishin]

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?

[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/showdoc.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.. :confused: 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...:rolleyes:

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.

:stick: no need to call someone an idiot :slapass: 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:banana::banana::banana::banana: 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 :nono:

*NO ABUSING*

[Cooper]

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

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.


:poke: :poke:

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$ ^^


:poke: :poke:

ich will remind you , when first reviews pop up :)


@mine: leicht verwirrt ^^ denglisch :D


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 :D

[FischOderAal]

ah, Poncho. what do you say about this article?
they used the Gigabyte P965-DQ6 with i965.
http://www.tweaktown.com/articles/914/6/page_6_benchmarks_super_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.