Just a guess, same reason old bioses are limited to 127.5 GB: 32 8 bit registers. Sensor stores data in a 32 8 bit register, gives 256 possible values,  which is mapped from 0C to 128C with 0.5C precision, and perhaps one reserved for error state or whatever. You would probably get the same value if you pulled a power cord from a running GPU, but Id recommend not testing that theory.

Precisely. I think some people misunderstand "hopping". Mining at random intervals, or only during day or whatever, is not hopping and "hop proof" payout schemes do not penalize this behavior. Quite on the contrary, even if you mine at random times, you would be gaining substantially more with PPLNS or DGM compared to proportional.

There is no upper limit to how long it can take to mine a block. He's at what, 7M now? Thats bad luck, but not extraordinarily so.  Assuming it is bad luck and not bad programming, thats always the worry with a new pool and a long first block.

Anyway, I dont see the point of continuing this. It would make more sense to just pay it out and switch to PPLNS asap, so perhaps people want to join without being paid 120% PPS.

I thought you had 50GH yourself, you could use that for testing, it would cost you nothing. If needed, rent more for short times at gpumax.  Why would you instead rent at 115% 24/7 for weeks on end when you havent even got a decent front end running, when you already know you are having stability and high stales issues that you cant fix for now? Any sane person would attempt to fix it first, and then test again for a short time. Not keep throwing money out of the window to test something you already know is no good - and incomplete.

Moreover, you never mentioned kickstarting a new pool in your OP. You blabbered something about avoiding exchanges and not getting red flagged by your bank, selling coins locally etc. Remember?
So when did you change your mind?

As for the flying spaghetti monster; perhaps an apt analogy, except you running a pool now rather than hopping is about as credible as the flying spaghetti monster.

Anyway, the point has been made. Its clear you are not going to just say the truth so,
whatever.

Looks like someone visited us or rented some hashing power:

Latest Shifts
2012-02-09 22:59
Completed    Ghps    Your Ghps    Your score
2012-02-09 22:26   181.778   1.695   0.93%
2012-02-09 21:32   162.272   1.695   1.04%
2012-02-09 20:30   133.778   1.666   1.25%
2012-02-09 19:16   135.422   1.665   1.23%
2012-02-09 18:03   136.809   1.692   1.24%
2012-02-09 16:51   139.309   1.661   1.19%
2012-02-09 15:40   139.454   1.757   1.26%
2012-02-09 14:29   136.595   1.707   1.25%
2012-02-09 13:16   135.904   1.651   1.21%
2012-02-09 12:03   134.322   1.745   1.30%

Block was found (YAY, well done Turbor), and now its back to 120GH. Its like reverse hopping .

Were you really?

Here we have a known hopper suddenly buying hashrate at well above market rates.
No credible explanation is given, only a demonstrable (and now half retracted) lie.
"pool" started around the same time as a new major hopper registers on bitclockers, that hops their pool at roughly the same hashrate as clipse's "pool".
Posting actively on github/bithopper about problems with too many connections, bithopper crashing at the exact same time as members of this "pool" report downtime and connection problems.
Posting on github that the server running bithopper is in a datacenter (IIRC in the UK). Would that make any sense if he only used it for his own farm in SA?
Reports of high stales that cant be fixed, much higher than youd expect on a proper pool, but most likely to be expected for redirecting hashes via bithopper
Delayed PPS payouts (enough time for hopping to pay off)

If it smells like a dog...

Most importantly; if you know how to hop, you have the infrastructure and software to hop, you have no moral objections to hopping (which btw, is fine by me), you get some 50 or whatever GH that cost you 115% PPS, why on earth would you decide to mine your own blocks at 100% with that, making a significant net loss, when you can achieve a reasonable profit using the hopping infrastructure already in place? This was a business venture, not a charity.

Anyway, this is getting pretty silly. Mostly because hardly anyone would care if Clipse had just said he uses it for hopping, and that includes yours truly. Anyone renting out their rigs for 115% most likely has no scruples about hopping.  Anyone wanting to hop but not knowing how should be happy. The only ones that have some reason to be unhappy are the proportional miners, but apparently even they dont seem to mind or they would have switched pools or their payout system long ago.

So why clipse keeps digging this hole for his credibility is beyond me.

lets hope there isnt a kink in the block finding algorithm

Of course that solves almost nothing (other than protection against the pools being down).
If 2 or 3 pools that have >50% combined decide to collude to launch a double spend attack in your scenario, thats still entirely possible.

Thats what I would do, but perhaps wait for someone else to chime in, as I have no personal experience with powered extenders.

No 2 chips are 100% equal, there is always a variance in static and dynamic leakage and therefore powerconsumption. That goes for the GPUs mostly, but the CPU or heck, even motherboard chipset could make the difference if you are sailing too close to the wind (and with 5 overlocked 5870s on unpowered extenders, you probably are).  Be happy it works on 4 of your rigs, and get some powered for the 5th .

Id actually consider retrofitting the others too if you can find the cables cheap enough, to reduce the chance of blowing up your motherboards. You may not need powered extenders on all cards, just a few per rig to reduce the load on the PCIe bus.

Anyway, thats just my 2 cents. Perhaps its something different entirely, but I doubt it.

This.
If i were to have any trust in these guys, id gladly offer 3GH for hopping at 100% PPS if the profits went to P2Pool bonusses.

Thats not my point. It would be worth checking if 5x925 MHz works on your troublesome rig or not. Power consumption of that is higher than 4x950 + 1x800, if it works, it would make power delivery issues with the motherboard far less likely. If 5x925 does not work, thats what I would betting on: too much power pulled from the PCIe bus. Are you using powered extenders?

Have you tried 5x900 or 5x925MHz?

You can get sub 65C on air easily with most cards, particularly 5850s like you mention, at least if room temperature is not summer hot. None of my 5850s run hotter than 65C and using an extreme air cooler (Spitfire) I even got one of my 5870s down to 34C (12C above room temperature) with virtually no noise. Granted, thats not cost effective either, but still cheaper than watercooling.

As to answer your question; yes, proper cooling should increase the longevity of your cards, but by how much is impossible to tell.
Even so the potential for a positive ROI just isnt there. Judging by reports on this forum, Im guestimating an average failure rate of less than 10% per year on HD 58x0 cards, probably even less than 5% if you ignore fan failures and mishandling. That gives you a ballpark estimate on how much ROI spending on watercooling could be worth in that regard; no more than 5-10% of the cost of the card even when generously assuming watercooling were to eliminate 100% of all failures.

In short: no, its not worth it from that perspective.

Bitminter client is indeed a lot slower than others for CPU mining, but no one cares, because CPU mining is so absurdly slow anyway, one has to wonder why you are even bothering. More over, it puts a lot of stress on the servers so there are even calls to ban cpu mining from pools, so dont expect your issue to be solved;  at least not by anything other than adding an opencl capable AMD videocard .

If you can not add a videocard, you would be wise to look into mining on alternative CPU friendly chains like Litecoin. Its not going to make you rich either, but you will make considerably more. You can trade your litecoins for bitcoins.

My point exactly. Pool hopping is a given, and I really have no "moral" issue with it.
Lying about what it is you do with other people's resources is something else though. In that sense I was pleased to see the OP was edited, but dismayed that Clipse now perpetuates his lie even after the cat is out of the bag.

Anyone voting for proportional either is hopping or doesnt understand hopping.

Those 115% PPS proxy pools by Clipse and Goat are hopping you and bitclockers.
I charted bitclockers stats,  did the math and calculated the effect:
https://bitcointalk.org/index.php?topic=61985.msg737310#msg737310

I havent done this for your pool yet, but I expect the results to be similar.

THe short version; anyone not hopping there but mining 24/7 is losing out on 20% revenue compared to the hoppers.

To think its "a wash" because they increase hashrate is silly. Each block only gives you 50BTC, if hoppers can increase their portion of that by 20%, who do you think loses ?

watercooling is too expensive IMO, unless its for 1 or 2 cards in your PC that you use everyday and you need quietness.
For farms its a non starter.

Im currently experimenting with submerged oil cooling as a cheaper alternative.

Good catch. Thats pretty convincing evidence in case anyone still doubted it.

It does support my point there is no driver limitation . From your link:

Whether or not those BIOS problems will be a factor using those backplanes is anyone's guess I suppose. The only way to found out for sure might be expensive .