Great review, but the conclusion doesn't make any sense: " It is quite possible that it would make more sense to go for a high-speed LGA1156 or LGA1366 processor instead of buying a new Sandy Bridge CPU at this point."

There is no point at all in getting LGA1156/1366 if you are building a system from scratch.

Looking forward to your 2500k/2600k review. I hope to see:

2500k/2600k @ 4.6-4.7ghz+
Core i7 870/950 @ 4.0-4.2ghz+
Phenom II 1100T @ 4.0ghz+

Also, can you please show the improvement in POST speeds between boards with conventional BIOS (i.e., Gigabyte P67) vs. boards with UEFI (i.e., Asus, MSI, Asrock)?

Finally, it would be great to have some tests with 1st generation Core i7 processors to see improvement in performance per clock (IPC) (i.e., Core i7 870 @ 4.0ghz vs. 2600k @ 4.0ghz and Core i5 760 @ 4.0ghz vs. 2500k @ 4.0ghz).

Thanks!

The great article, just two remarks:
1) Are you sure that in TrueCrypt test SB did use new AES instructions?
2) Do you plan to test Quick Synch?

Yeah. it would be great t osee some different generation comparison tests on same frequency. Even with a Core 2 Quad 9650, for example. Thanks.

Can someone please tell me if the SB integrated graphics are able to handle dual-DVI?

Specifically, can it show the same video output in Clone Mode? For example, can it display a movie on a computer monitor though the DVI output, while at the same time diplaying it on a TV set through HDMI?

----

Other comments: the Overclocking page says the base clock can be raised. This is not what all other reviews I have seen say. Only the Turbo clock is supposed to be adjustable. Also, the overclocking page says that frequency can be raised by 4 steps for non-K CPUs, but the example shows a 3.1 GHz Core i5-2400 processor being increased to 3.6 GHz (5 steps). How can this be?

I am confused...

It's a weird conclusion, really. If we bring the 2600K to the table, the advice for sticking to 1366 makes even less sense, as this CPU is faster than the 980X in most occasions, at a fraction of the cost.

It's interesting to extrapolate the results seen here to 8-core parts; now that would be insane CPUs. Given the low power consumption of the quad-cores, it's not impossible for Intel with some optimizations to the design to roll out a 12-core part as well - this have to be Xeons, though. I can't imagine regular users needing that much power right now.

If those rumors about the performance of 8-core Bulldozers is true, we may see those 8-core i7s sooner than expected, I guess. Interesting year ahead