<offended>Who the hell is calling me a whitehat?</offended>
On a more serious note, if you're going to put a hat on me, it'd be some shade of grey, but sure as fuck not white.

Broke GG1 (orphaning other miners + timewarp to drive difficulty to minimum) and "broke" SC1 (53 or so oversized tx bloating the chain and slowing things down, inadvertently triggered a bug/lack-of-optimization inherited from bitcoin that crashed a few nodes by consuming ~4GB of diskspace).
While were at it, also found a node-crashing bug in bitcoins script interpreter in late '10 and crashed all testnet nodes.

As for the accusations of me attacking I0C ... no clue where those come from.
I0C broke because the joker who originally modified the retargeting fucked it up, and it flat out stopped working after 2 weeks.
Then a patch from me fixed it (well... "fixed"... just ripped out the broken mess and replaced it with a modified SC1 retargeting).
Long after it got relaunched and everyone lost interest in it again, bitparking i0c exchange got hit with a doublespend, but that wasn't me.
I don't think anyone clearly admitted to that, iirc BEX hinted at having something to do with it... wouldn't be too sure, he also bragged about several other things I know are not true.

So in case you haven't noticed yet, I tend to break things for fun, not profit

Well, Gavin already replied and stated his personal opinion, and any other bitcoin devs with half a brain will keep mum and stay out of this whole mess (hint: they probably don't give a flying fuck).
So... thread done?

Yep, another case of not understanding how bitcoin works.
I've watched people come up with the same or similar bullshit "attacks" several dozen times in the last 18 months.
I'm pretty sure there's even a FAQ entry for it on the wiki.
But hey, you don't have to trust anyone, just read the whitepaper, read the source code, think for a bit, if you still don't believe it *try it*
... and please post after you realize how clueless you've made yourself look.

Anyone with a copy of the CLC blockchain and bitcoin-utils can easily verify that the bitcoin parent blocks used to attack CLC belong to eligius. Personal equipment my ass.

100% agreed.
If your system relies on a lack of assholes on the internet... good luck.
Btw, the same problem also exists for non-mergemined chains with miniscule hashrate right after launch, remember the massive FBX orphaning?
Anyone with a dozen decent CPUs could do it... and someone did it. whoops. Except no one stepped forward admitting to the attack, so instead of 20 pages of forum drama spread over a dozen threads we got a bunch of "aww, sucks" posts and people moved on.

Well... yeah. If you happen to find a ATX psu that's just barely within spec, it's outside of the nominal supply range of a S6. By 15mV.
ATX spec also requires a 1A min load on each of 3.3V, 5V and 12V, so below that regulation can be bad (a real problem with group-regulated PSU designs).
With modern high-efficiency psus generating their 5V and 3.3V with step-downs that issue seems to have pretty much vanished, some mV ripple at a few 100kHz+harmonics, nice and well behaved load step response even down to 0 load.
As for seeing nasties on a 'scope ... are you measuring on a unloaded/unterminated line? Try with 10uF || 330R, or you'll be mostly measuring weakly coupled interference, transmission line effects/line resonating with your probe capacitance.
But then I've only closely looked at 2 zippy and 2 superflower units, all of em 80+ gold or better and all of them good enough that I decided the added cost+complexity of a 3.3V step-down for each 8*LX150 board wasn't worth it.
Had more issues with vccint droop, on my boards a 3-phase DC/DC was at one end of a 190x85mm board... close to a 50mV drop to the furthest pair of LX150s. Salvaged by adjusting the converter up to 1230mV and adding lots of jumpers. rev1.1 moved the converter to the center of the board and rev2 changed the board geometry completely (2 rows of 4 instead of 4 rows of 2).

Tight coupling between UI and core (somewhat reduced with 0.5), Horrible OO design in parts, data members being public in lots of places instead of having accessors, sometimes outright WTF-worthy approaches to doing things. I wouldn't say incompetent, just lazy/messy.

I don't get what the huge outrage is about.
Author of FOSS miner offers to work on improved 7970 kernel if he gets enough donations towards buying a 7970.
That's pretty much the same way you get your personal pet hardware supported by FOSS firmware. Guess next you'll be boycotting dd-wrt and the likes.

... says who? According to spec it should work fine, and reality over here agrees. 54 S6s happily running with vccio and vccaux fed from atx3.3.

In case you're wondering why the SSE2 version sucks on K8 and K10 ... reason is rather simple.
the salsa20 function is a long string of data dependent 4*32-bit vector integer operations (i.e. output of one operation is used as input to the next).
And the execution latencies for the most used instructions in the salsa20 core (shift r/l immediate, add, xor) are all 2 clocks on K8/K10, all 1 clock on Atom/Core/Core2/Nehalem/SB.
End result ... sse2 salsa20 needs roughly twice the clocks/round on AMD compared to any modern intel.

Well, it changes voltage for the wrong VID on 5970s by default, on 5970s (and iirc ref 5870s) VID 3 is for performance mode, on ref 5850s it's VID 2.

Yup, sure looks like a used card.
That TIM on the GPUs isn't original; all ref 5970s I pulled the fansink off of had phase-change pads, not goop.
Also, good idea with the dead phase. Check VRM temps under load, if one of the 3 phases is a lot (20C+) cooler than the others you got a dead phase; the card will generally still run but unstable and OC like crap.
If that's not the case... idiot who replaced the GPU TIM didn't replace the thermal pads for the VRMs. Most important ones are the one for the core VRMs. the long one under the fan and the smaller of the strips above the vapor chamber. inductor, ram and pcie bridge pads aren't really that important. Used a bunch of a 1mm sheet of 3M 5591 I had lying around. worked like a charm.
In linux a slightly modified radeonvolt works for reading VRM temps and overvolting 5970s.
As for the holes in the paste... bet the monkey who applied the paste touched the die and didn't degrease it before applying new paste. that generally ends up looking *exactly* like this.

Hard to tell from pics, but ferric oxides (aka rust) def. can cause such a brownish tint.
Steel dust from grinding would be a good source. Did you have the loop open while you were dremeling the case?

If you don't have a irrational phobia of mating Sn and Au plated surfaces, 45750-1112 should be the right one.
I used 45750-1212 - same thing, just gold plated and like 5 times more expensive.

Enigma81: You sure?
13A is only when loaded with mini-fit HCS contacts, which is only required by the spec for PCIe 8-pin connectors (and possibly EPS 8-pin, too lazy to check), but normal atx 20/24 pin doesn't require em.
So we have normal mini-fit contacts, wire-to-board, for AWG16 or AWG18 wire, brass contacts are rated for 9A, bronze contacts 8A. derated down to 6A repective 5A depending on # of loaded circuits in housing, board layout, ... (read the fine spec sheet for the contacts, it's freely available from molex).
And that's assuming those are real mini-fit contacts, and not some cheap sub-spec lookalike... fat chance.
As for desoldering coarse stuff, *love* my Weller DS100.

Back on topic.
Burnt the 12Vs on a atx 24-pin running quad 5970s on a MSI 790FX-GD70. Replaced the connector on the board and the plug housing, replaced the 12V contacts with mini-fit HCS series. Preemptively replaced the 2 12V contacts in 5 identical other boxes (one had the plug already partially melted around one pin). No further burnt 12V lines operating 24/7 for over 6 months.
Other stuff:
Dead fans (both the well-known bearing failure) on 2 reference 5970 coolers.
4 Sapphire nonref 5770s with dead fans.
>20 Sapphire xtreme 5850s with dead fans.
Fried the VRMs of the 2ndary GPU on 2 5970s with AC Accelero coolers.
Coolermaster 650W 80+ bronze blew up running 5970+2*5770 (12V output filter coil burnt).
3 Andyson-based 1kW 80+ gold PSUs blew up, each running only 3 5970s at ~880W at the plug (secondary side synchronous rectifier FETs failed).
Box full of dead Scythe UK3K 120x120x38 mm fans, all bearing failure. Replaced with deltas... buy cheap, buy twice.

and 370 / 905 * 940 = . Oh.

Same as bitcoin, difficulty 1 = about 2**32 hashes average per block.

Not sure on power, someone bored enough to synthesize a design and run power estimations?
Don't get fooled by xilinx S6 perf/power, those are the worst of all 45/40nm FPGAs by a long shot; we pretty much only use them because they're cheap-ish and readily available in small qty.

Those packages look like 1mm grid thermally enhanced flipchip BGA ... guesstimating by the 0.1" headers ... 29x29mm FF780
so if these are SIII, pretty much "could be anything from a L50 to a E260"
One thing kinda throwing me off... if those are HCs, 1.1V core doesn't make sense. HCII is 1.2V and HCIII and IV are 0.9V.
As for "what SIII would you need for that"... depends on how much time you spend optimizing, first guess would be 2 unrolled engines with 1 pipeline stage per round, each running @ 250MHz... gut feeling says a L150.
So, random guesses:
a) these are HCIIIs, they prototyped with SIII and forgot to update the label for vcore.
or
b) only SIII based prototypes exist, once they get enough pre-orders HCIIIs get ordered (about 8wk if you rush assembly... possible.)
or
c) they somehow got a whole bunch of large SIIIs really cheap.
or
d) something entirely different.

ngzhang: Altera SIII is 1.1V core.