Okay. So Intel is making some strides with their IGPs. Not that it was benchmarked, but I bet you could build a decently priced WoW box off this. And the performance would be acceptable, I think.

And btw, at first glance I would've figured that board to be a Gigabyte. The colors! The colors!

AMD's SATA performance is stupid.

I'd take a G45 mobo in a second. Good performance where it counts, and no illusions of having gaming performance for free.

Thanks for the article TR.

I'd just like to say that this was an excellent review and the photography was superb as usual. Thanks!

Interesting side not:
I just built some budget computers for my friends. When i first started out, I intended to use a intel CPU. But after looking at the choices with integrated graphics; I ended up going with an AMD CPU just so i could use 780g.
I think its funny that the CPU didnt decide what platform i chose, but the motherboard did, which is usually just a peripheral choice for me.

Geoff are you using a Vista install that has SP1 slipstreamed? Because the new Vista DVD's with SP1 slipstreamed don't give any issues with AMD AHCI anymore.
I'm using SB700 AHCI with Vista SP1 without any problems. Microsoft implemented the fix the offered before into the SP1 install.

Darn, no factory oc'd IGPs? I'm afraid this article's comments won't be nearly as numerous or intense as the GTX260 Core 216's :( Damage keeps all the good stuff for himself eh?
=====
Joking aside, nice overview. No mention of the G45 delay though? It was supposed to be out by the end of August or earlier originally iirc.

The only other thing I can say atm is that some board makers continue to get big FAILS on certain simple layout things. Blocked SATA ports is not acceptable, there are right-angle ports out there or at least put them inline with PCI slots, Asus has been bad with this on various boards for a long time you'd think they'd know better as the market leader. And wow! look at the ATX power connector on the Asus! MSI gets a downrating for putting the PCIe 1x slot next to the PCIe 16x in aqdditoina to the CPU power header. Gigabyte even gets a semi-fail here, putting a card over the IGP heatsink seems like a bad idea to me, even with the smaller process of P/G45 I bet that NB gets hot.

Once again, Intel IGP is a shining beacon for all things that suck. Sure, they're not in the discrete GPU business so they don't have that IP to borrow from, but with all their millions of BushBucks, this is the best they can do?

This is the first time I've overclocked our E2180, and I was surprised to see it effortlessly sail up to 3.1GHz without so much as a voltage tweak.

I'm surprised you found that surprising. It's well known the 2180 can often do 3+ GHz stock voltage. Running one at home on a G31.

Well, it's good to see that they are atleast improving on non-gaming usage scenarios. And after using a 915GM for a few years, the fact that this one will actually RUN games, even at like 5FPS is still an improvement in my books.

No HD decode image quality analysis?

Modern IGPs are good for HTPCs, not gaming. When considering that a Radeon HD46xx can be had for under $100 these days, there really is no place for IGPs in gaming.

However, not everyone is a gamer. I think an IGP makes a lot of sense for a media box, especially one in a small enclosure by the TV. I'm glad that TR didn't seem to have as many driver issues as Anandtech did when testing the G45 - hopefully, this means intel is finally fixing their bugbear of lousy graphics drivers.

Still, with Nehalem and X58 due in a month, it's hard to get excited about the G45. If it had come out 3-6 months ago, then we'd be talking. Hopefully, the G/P55 will be a contender - my guess is that it will be paired with the lower-spec 2-channel Core i7 due next year.