I'm getting a weird problem that I've never seen before. I haven't changed any drivers or sdk or anything, it just started happening out of the blue.

Someone thought it might have something to do with bip16 or something to do with local block detection (?) and cgminer had this problem at some point?
I really dont know, i'm just posting what someone else though. Can I get a fix please? I'm mining on eligius.
The miner continues to submit "shares", but they are all rejected with the reject reason: stale-prevblk. when another LP hits, it doesnt spam anymore and it submits valid shares

MSI sure seems to be the flavor of the year. Lots of great quality products without the extra flair and price tag (such as Asus). Even their Big Gang Xpower II product beats Asus Rampage in price and capabilities.

But yeah, I'll vouch for this board, even if I don't personally own it. Make sure to get a good deal on it.

The latest is not always the greatest.

I suggest using 11.12 or 12.1 driver for using 5xxx or 6xxx cards, and using 2.1 SDK for 5xxx cards and 2.4 or 2.5 for using 6xxx cards. Guiminer, and some other miners/frontends, will not detect 2.1 unless 2.4/2.5/2.6 (2.6 is not recommended, much slower) is also present. Verify your cards, SDK, and driver with GPU Caps viewer.

Hmm thats a full 8 mhash more than what I would project for 1120 SP's at 850MHz.. Whats your OS/driver/SDK/miner/kernel/settings?

I have 2 multi GPU PC's mining, none of them use crossfire and they all mine fine.

now thats a lowball if I ever saw one. 175 per card = lol

the f argument for diablominer is pretty much what you want. A higher number = less hashes. it's effectively the "framerate" or the target for how many times/second a certain set of nonces are hashed. Lower framerate = more nonces per run = gpu works longer. Try a high number such as 240, or even higher until you are happy with your hashrate and gpu usage.

Wow, why are you using such high memory clocks? That is sucking a huge amount of electricity and causing you to have higher temperatures and losing money. Underclock your memory and stop using shitty 2.6 SDK. Use 2.1. I get 294 mhash @ 900/355 using 2.1 SDK.

Wel if this is about ebtter paying pools, mine at Eligius. It's 0% fee and PPS. Don't even need to setup any logins or anything, just point your miners at http://<your address here>:x@mining.eligius.st:8337. Payout is automatically queued (not guaranteed immediately, but queued) around 0.67 BTC or a week after your last submitted share. Stats for your "login" can be seen at http://eligius.st/~artefact2/7/<your address here>. Mine as an example is http://eligius.st/~artefact2/7/1CxcPP8FVktppy4PHTYJKnZFqQeyZ3jArb . Any of the huge variances you may see is a result me manually adjusting GPU's clocks/volts to agree with the warming weather.

You cannot crossfire a 5xxx with a 6xxx. It is impossible. You can crossfire a 5xxx to a 5xxx and a 6xxx to a 6xxx though, but your speed will be limited to your slowest card times the number of cards in crossfire.

12.3 had a bug which prevented some software from overclocking.
also make sure you are using sdk 2.1, its the fastest for 5xxx (2.4/2.5 is -ok-, dont use 2.6)

Bah. I use a 5870 in my gaming PC, and I'm currently troubleshooting SSD nonsense, which requires me to unplug the card to get to the SATA ports. That would really get troublesome with restrictive tubing, unless there is a way to unhook pieces within a loop without causing it to leak horribly.

Otherwise I would be interested

1 share ≈ 2^32 hashes or 4294.967296 megahashes
60*60*24 seconds in a day = 86,400 seconds/day
1010MH/sec = 87,264,000 Megahashes/day

Shares/day = (MHash/day) / (Megahashes/share)
Shares/day = 87264000 / 4294.967296
Shares/day ≈ 20317.73
Payout/day = 20317.73 * 0.00003 Bitcoins/share
Payout/day = 0.60953199 Bitcoins

Feel free to tip.
1BXNapSrzTSfmqSD63e49eh2JvDkeotCxt

Could very well be a bad "extender". I threw out all 8 of my PCI-E 16x>16x and 1x>16x because they would either not detect the card, or it would be intermittent. Too much bending broke some of the many solder points and I don't have any solder skills. Now I use 1x>1x with open-ended slot and pci-e presence pin short and they all work like a charm.

Source?  
I have no clue how to "properly" bid on ebay with automated snipes. I was looking at $0.075 per SP (means a 5750 for $54, 5870 for $120, etc) shipped, and I'm just blanket bidding on everything on ebay.

I also don't exactly have a deep wallet as I'm just a lowly grocery store clerk working part time. 2.5 Ghash have supplemented my income substantially.

Can also try a significantly lower f number. A higher f number means more responsiveness of the desktop at the cost of less GPU time dedicated to actual mining. A lower f number will enforce "bigger" nonce ranges per kernel run, causing more desktop lag but more dedicated gpu time to mining. You can think of f as the framerate, but a lower framerate means better mhash.

I'm super jelly that he can just use fresh cold tap water to cool everything instead of us normal people that need to screw around with a radiator, a pump, and a reservoir. Must be soooooo nice having free water. Dont even need air conditioning! All the heat goes down the drain!

Are you sure you used 4.3.0? I can go to 2GHz+ or lower than 100MHz on my 5870 on core, and the same for memory. My volts can also go from 450mv to 1300mv. If you are maxing/minning a specific gpu/memory clock, just hit apply and the slider should make room to go even lower/higher. There are no "hidden" settings or config or anything like that. You just set and go.

Your CPU and mobo is incompatible. Socket FM1 is only for the APU's (a4/a6/a8) that have integrated gpu. You need an AM3/AM3+ socket mobo or a socket FM1 APU.

Your card is in UVD mode, which automatically enables when you are using your hardware to decode video. This includes almost any flash popup on websites (youtube, newsground games, etc). Close your browsers and videos or disable hardware acceleration.