Great article Scott.

You guys very badly needed to test the existing i945P/i945G boards along with the P/Q/G965 chipsets, just to compare the current generation with the previous one.

(Yes, there are plenty of i945P/G boards that support VRM 11 and Core 2 Duo. I know several people who are running that combination -- prefer it, even, because of the lack of Linux support for the third-party PATA chipsets found on the x965 boards.)

I'd also like to see a comparison of the C-2 stepping P/Q/G965 boards against the C-1 stepping. Apparently there have been numerous improvements in this latest revision.

Some fanboys still stubbornly cling to their favorite underdog, but most enthusiasts have seen the light and are looking at Core 2 for their next upgrade or system build.

Heh...;) That's a pretty amusing definition of "fan boy" and "enthusiast"...;)

At home I'm still running an AMD 4000+ on an AGP motherboard with an ATi R420 AGP gpu and still consider myself an "enthusiast" even though I have yet to jump on the X2, PCIe, or R5xx bandwagons. I haven't "upgraded" to the "nifty new stuff" for the simple reason that my "nifty old stuff" is still doing a great job for me and I have no complaints. If I am "stubbornly clinging" to anything it's only good sense I'm clinging to, imo...;)

By your definition an "enthusiast" would seem more akin to a "dunce"...;) Seriously, are you going to write the same slant in 6-8 months when AMD is shipping the K8L, and are you then going to be chastising Core 2 owners to again "see the light" and dump their months-old Core2 systems for K8L systems just so that they can be considered "enthusiasts"? I very much doubt it.

To me the only "fanboys" remaining are those who "stubbornly cling" to the idea that Intel reigns supreme, now and forevermore....;) I mean, history has proven otherwise, hasn't it? You might want to rethink some of your characterizations. I think that "stubbornly clinging" to good sense is an admirable trait.

Heh, funny to see every website nowadays whore itself out to digg.

Not that I don't love TR, just an interesting trend.

What will the chipset market look like next spring? Anything important on the horizon?

Digg it, please.

http://digg.com/hardware/Core_2_compatible_chipsets_compared

Bummer no 946 or 963. I see alot of board out there using the 946...or a 945...that are Core 2 compatible.

It would have been nice to see if a 946 is any slower than a 965...in real world testing.

They are cheaper.

Techreport needs to review the Intel mATX DG965OT motherboard! Please? :)

did anyone mention 965 problems to boot the first time with memory modules which require more than 1.8V?

I think this is a critical problem for "ease of use" on that platform, and quite a few people have had serious issues on that front.

I am definitely looking to upgrade soon in the future (still running an dual AMD mp1800+ w/ 8500LE system)

My choice presently is also between 975X and P/G965, but for an odd reason: Right now the only SFF system that supports Core 2 that I can find is Shuttle's SD37p2. Since that system is 975X based, and has some features that i'll never use (crossfire) it is quite expensive.

$440 for a SFF MB/PS/Case

or

$440 for a mid tower MB/PS/Case?
Lian-Li PC-V1000A Plus II
Coolermaster RS550
Intel BOXDG965WHMKR (G965/ICH8R)

I'm trying to figure out how the 570 is an upgrade.

I still have rather have P965 for overclocking and 975X for ECC support and having more PCIe lanes then P965.

Nforce 4.5 SLI for Intel (NF5 570 SLI 1.0) looks good on paper, but the higher power consumption and some of the known quarks of previous Nforce chipsets when compaired to Intel chipsets makes it a no-brainer.

Some fanboys still stubbornly cling to their favorite underdog Even the site's authors throw flames once in a while. Ah well, no harm done.

That first photo in the article looks like something you'd see at Tom's Hardware for some reason. Tom's always does those "compositions."

When will we see a 975x uATX motherboard already?!

When you list results for scaling from one card to two cards in both crossfire and sli perhaps the use of percentages would help readers better understand the difference rather then just looking at raw numbers?

Just a small comment on the PCI-E gig-e test as well. If you look at theoretical bandwidth a gigabit controller can't saturate a x1 slot. Catch me if I'm wrong but a single x1 slot offers theoretical 250MB/s one way or 500MB/s both ways, by contrast if you convert 1,000Gb/s to MB you get about 125MB/s (1,000 / 8), well under the theoretical, 250.

Maybe the use of a cheap x1 SATA raid card would help saturate the bus to the point of deferintiating the boards?

Why isn't Crossfire available for the 975 when it is for the 965?

Good write up, wish it came sooner though.

Just curious, why does the P5B Deluxe use an old beta bios? It dont even use the 0507, or even the latest 0614 retail bios. 0505 is just very very old and not retail and contains alot of flaws.

http://techreport.com/forums/viewtopic.php?t=43787

And since the P965 is a very new chipset, each BIOS update includes quite abit of new things and changes. The alot newer BIOS for the P5N32-SLI SE Deluxe is on 0401, unless its more than a typo, but not this one? P5W DH Deluxe is also 1305 for latest BIOS.

The chipset drivers are also not the latest, and its listed wrong for the 570 board.

I´m alittle dissapointed. And I figure you still use the B1 engineering samples instead of retail B2 even tho Intel updated for a reason just before launch.

So nice review but useless, now update it and try again. :(

Nice article, Geoff.

Anandtech previously ran a series of articles on C2D memory performance on several platforms which included testing at varying memory speeds. Interesting were that,

1) They found no significant performance differences in real applications between any platforms or memory speeds, even DDR1-355 on i865.

2) Within the probably meaningless 1-5% margin of performance difference, the 965P performed consistently poorly with DDR2-533 memory.

Their testing was rather limited in that the processor was a E6300 at stock speed, and the video card used was a 7600GT to allow direct comparison with a DDR1/AGP platform.

A cross platform test with components as an enthusiast would actually use (E6300/6400 bumped up to about 3ghz, 7900GS or X1900) including varying memory speeds would be interesting.

If the P965 1:1 memory divider is slow (in the niggling 1-5% sense) at stock speed and DDR2-533, will it choke with a 400mhz FSB and DDR2-800 as well?

Would be interesting to see how Crossfire works in a 16x/4x setup vs a 8/8 setup sometime in the future.