Results: Applications

nice to see a huge boost in productivity/non gaming applications .

Nice to see a huge boost in gaming/non productive applications as well.

Good article guys, the screenshot gallery is great, thanks for that! I'm a modest OCer these days so sub-4GHz overclocks w/ little to no voltage increase is exactly what I'm interested in and the screenies are useful as a reference. Even better, you used one of the 3 main mobos I've been considering.

Question about the MSI GD65: Patrick/Achim did you guys try using any Channel Well based PSUs and if so did you experience any of the problems Atech mentioned in their review? (scroll down to the 4th paragraph) I recently purchased a Corsair 750hx and have noticed a fair number of comments not only in that linked article but also in places like newegg etc. The OC Genie issue doesn't bother me much but just hearing that there's any sort of problem at all between MSI boards and CWT-based PSUs has me a lil' sketchy till it's (presumably) fixed in a BIOS update. The Anandtech article mentioned they had the exact same problem w/ the GD80 so till MSI sorts their P55 problems out I'm on the fence about going w/ my first choice, the GD65 (LOVE that board's layout), or instead switching to the Asus/Gigbyte offerings.

Dunno, I think the i7 920 is still better.

i7 920 should be better, it costs an extra $100+ for an x58 mobo that is equivalent to its p55 counterpart. Yet, at stock speeds i7 920 loses to 860 in nearly every app and when both cpus are OC'd, it's still close enough that most people will ask themselves if it's worth the extra $$ for x58. i7 920's main advantages are:

~ dual-CPU workstations
~ high-end triple/quad GPU gaming setups that will tax PCIe throughput
~ easy (relative to p55) 4GHz+ overclocks
~ bandwidth intensive tasks that make use of the 3rd RAM channel.
~ nearly a year headstart to stabilize drivers and mature the x58 mobos

What seems to be more and more in doubt is if x58 is still better at:
~ stock/low voltage overclocks w/ little tweaking required to hit ~3.5-3.8GHz

No. The over-speed protection limits you from going past a certain power threshold. EE CPU's have unlocked multipliers, making overclocking extremely simplified.

Originally, you could only disable over-speed protection on the EE i7. Shortly after release, Intel changed their minds for whatever reason (Likely very bad publicity), and fixed that issue. Now it's entirely up to motherboard manf's to decide to include the feature to disable it.

And you're totally right, there's not much of anything worth the $700 price premium. The EE processors are more for people that either have more cash than sense or who want the fastest available, but who don't want to inherit the risks from overclocking.

Wow, the performance per watt of the turbo was a little too hogh, but the app boosts are incredible, gaining over %33 productivity. Thats a great o/c to an already great cpu. Good article Pat & Ach.

This is great. It would be nice to see these numbers side-by-side with some other processors (AMD?) for comparison purposes though.

I agree with you... but of course I have a i7 920

I like the thought of upgrading to the Intel Hex core next year over the current benefits of the i5 750 cost to performance.

Good article. I'd like to to see come comparisons with the i7 920/940 though.

Is there anything else that needs to be changed for overclocking without voltage modifications?Is it possoble to say set the 133 base clock to 160 and be stable without doing anything else?

Interesting article. I'm a little suprised they didn't show any comparisons with the i7 920, but I guess they wanted to focus on LGA 1156.

I wish you throw in a PII X4 at similar OCed speed just to show scaling. A 955 UC/OC to 2.66, 2.8, 4.0 & 4.2 will be very interesting to see as many people are saying PII @ 4.0GHz = i5 @ 4.0GHz which I think is not true.

An i5 750 @ 2.66 (stock) is already faster than a PII 965 @ 3.4 (stock). What makes you think it's going to be any closer when you push the i5 to 4.0?

Patrick/Achim, there's something I don't understand re your choice
of comparison CPU. The article states early on that the 870 doesn't
make any sense from a cost perspective, so why is it in the
graphs? Surely you should have compared to the 860? And indeed,
seeing how an 860 oc's is exactly what I would have thought most
people would want to know, since that is the most comparable CPU
to the 920. Paired into graphs with 750 results, it would be
genuinely useful to see how a 750 compares to an 860 on P55, ie.
whether the extra cost of the 860 is worthwhile (offhand, I had
assumed it would be) and by that I mean via including both stock
and best-oc 860 results.

Thus, can you please check with an 860 aswell? For most enthusiasts,
the 870 is not relevant. Likewise, how high can an 860 oc without
a voltage increase on that board? Does its higher base clock help?

Ian.

I'm going to get bashed for this but amd and intel chips do scale a bit different, but I do think this chip is sweet, and is worth the money. I would never be able to argue that my p920 could beat this chip. I was at a little overclocking comp a while ago and the chip that beat all the i7 and p2 was one of those big fat 775 chips clocked to high hell lol. so I don't know what to do anymore.

if you guys are going to do an overclocking article, at least try to not confuse readers with voltage terminology. You mentioned increasing the PCH at 4.0GHz, then show the PCH in the 1.8-1.9v range??? The PCH is the southbridge, and doesn't really need to be adjusted beyond the nominal 1.05v. The voltage range you are showing is that for the CPU PLL (I know they call it PCH_1.8v at MSI). Also, 1.45v for 4.0GHz, yep definitely a MSI board, that's horrible.

very interesting, though i would really like to see how far can i7 920 be overclocked and being stable.

Do we know if this particular board has a foxconn socket?