K404

What causes it, what fixes it? One of my cards does it from 1325MHz! and i've tried two shims, both my pots plus remounts. Contact looks fine. I thought it was heat/ bad contact to start with.

Whoa! That's.... unexpected without SB-E! Congrats! Guess I gotta install 3D11 on Z77 now

I saw this on KPC, from Ronaldo and PJ Is this only a "quirk" of some cards, or all? Only one of my cards appears to do this. I guess I could use it as card 2 in CF.... but i'd rather know what was going on in the bench if i'm benching it alone. Thanks

Has anyone who's stuck with this problem tried a Sandy bridge K chip to see if that clears the multi? Gamer suggested that by accident, looks like it didn't work for him though

That doesn't rule out the bug- it's a bug with Ivy Bridge Interestingly..... that's not a bad idea, i'll point that out in the IB thread, with reference to you

You say you're stuck at x39..... you sure it's not a form of the multiplier bug that others are experiencing?

A superPi time of 6.28 is 5GHz.

Sorry to hear bro You will not be the last person to have such bad luck though

Christian "Never forgets a meme" Ney

In honour of this: I give you this:

If it's happened to most mobo brands, maybe it's an Intel errata

If guys are taking great CPUs and changing the TIM..... and getting worse MHz, it makes me think of the fable of "The Golden Goose"

Ever heard of the internet? :D It means you can do stuff like that from almost ANYWHERE in the world

If you work for MSI, you could at least go and help Jon at Computex

That answer is one billion times better than the number of times I hear "sure, i'll send you one" only for nothing to appear at my doorstep. (not MSI, never asked them for anything)

I've never lied to vendors. If they send me stuff, i'm going to be honest in my public review. The stuff is probably gonna die. I am very unlikely to take any world records because that means going up against guys who have access to hundreds of CPUs and/or cheaper LN2. (I leave out the skill part) The worst thing a company can do to me is promise that my tomorrow is the same as my today and yesterday- no support. I've been doing this for 8 years. It clearly doesn't keep me awake at night. Hello MSI. While keeping all the above in mind, please can I have Afterburner Extreme?

How is NDA going to be enforced?? All a user has to do is cover up the username in AB. Plus, AB works on non-MSI cards too! *shock* Much like Asus allowing Intel CPUs to be fed with 2vCore... it ain't Asus problem if the CPU dies. Why would MSI care if I set fire to a Sapphire card?

There are plenty of "+" on reference PCB, just rebranded. If you have any other 128SP/ 256-bit G92 cards, bench Vantage (of course at the same MHz.) You will see what I mean about the ref being inefficient. I've tried a load of ref BIOS, they're all the same. Think I even tried a non-ref BIOS on the ref card and performance was still in the toilet.

It's quite optimistic to think that because you have 4 chips of varying quality, the silicon has another GHz in it....

Please do not ruin your reputation with lies.

Isn't it a common problem that AMD efficiency bombs at high HT? Example@ Sempron wPrime runs are at 6.4GHz..... PCMark05 runs are at ~4.5GHz. Of course....there are still guys running massive MHz, so there is still an efficiency comparison

I thought CPU-Z was updated because of how AMD chips behave? About the video: if it's stable enough to pause and film, it ain't on the edge/ it's in no danger of crashing. If I was at 9GHz, I sure as hell wouldn't be so chilled out Or.... the BS answer: the psychology is wrong. Break the CPU MHz WR by 500MHz and just react with "yea, whatever. I'll stop and film it" LOL really?

When I froze mine, the vGDDR circuit cut out. Came back to life after 2 days. My card was utterly average. Post up high-res pictures and I can help you out, sure

I kinda wondered..... but the feeling of 8.8GHz was nice while it lasted. 9062MHz "that easy" at <1.90v. Sorry, it's getting a little bit silly now. http://valid.canardpc.com/show_oc.php?id=2385966