Vcore of .3V to .8V would be better, IMHO, spawning efficiencies of ~.06W/GH/s and ~.15W/GH/s respectively.
Guestimates per the vid's

Patiently awaiting code release
Signed up for AirDrop.

Early Cisco IOS's were heavily based on Linux.
When NT was released NT had dynamic memory allocation/deallocation, Linux did not.
To me DOS stole more from CP/M & MP/M than the PDP 11 OS.
Much of the NT design team was "lured" away from DEC, so it makes sense that NT has "under the hood" similarities to RT-11.

Core Dumped Blues

Well, my terminal's locked up, and I ain't got any mail,
And I can't recall the last time that my program didn't fail;
I've got stacks in my structs, I've got views in my queues,
I've got them : Segmentation violation -- Core dumped blues.

If you think that it's nice that you get what you C,
Then go : illogical statement with your whole family,
'Cause the Supreme Court ain't the only place with : Bus error views.
I've got them : Segmentation violation -- Core dumped blues.

On a PDP-11, life should be a breeze,
But with VAXen in the house even magnetic tapes would freeze.
Now you might think that unlike VAXen I'd know who I abuse,
I've got them : Segmentation violation -- Core dumped blues.

Mighty impressive.
AND
Congrats on your first Linux build.

Most build Linux 4-6 times the first time, there are just SOOOOOO many options cuz' Linux is so configurable.
What distro did ya' start with?
ALL our servers run Centos 6, which may not be optimum for what your doing.

Sticks came in yesterday, Muy Bueno.
Took me a couple of hours to find the support thread and Novak's brilliant work.
Now hashing @ Eligius @ 200 MHZ, ~10Gh/s each.
Them LEDs are BRIGHT, Tooooooo F'ing cool (I'm such a redneck).

"... but how else do you stop companies from doing this if you require their service for your own business?"

Make it yourself, and if you can't do it where you are, move to where you can.

I moved from Dover, DE in 97-98 partially because power was becoming SOOOOO expensive (I read the writing on the wall) and the city wouldn't let me do wind/solar/bio-fuels at my house on Division Street.
First ~5-7 years here in MN, I made ALL my own power (and did not live like Grizzly Adams).
In '01 I got involved in county government and wrote the code and zoning for wind/solar (it's easy to win, if you write the rules).
In '05 I had'em (my local corrupt co-op) run in power, cuz' it was cheap.
It's tripled in cost since then.
With the cost of solar being so cheap, I'm re-visiting self generation.
And since I wrote the code, I know what I can do and what I can't.

If I can't get efficient chips, I'll fight this battle from the supply side versus the consumption side.
Besides, I remember how good it felt to NOT have a re-occurring monthly electrical expense.

During the school year, every Friday I mentor the local 7-12 grade science classes with my "radical" ideas.
Presenting strong economic reasons for doing what I do/believe (the proof is in the pudding, so to speak).
Cuz' it's hard to argue with numbers in black'n'white (Reality What a Concept -- Robin Williams).

What I'm trying to say is: it's not enough to just do it, one has to bank it forward to the next generations.
I endeavor to teach them they have options, to critically think about what's possible and look for better solutions.

Even the USB stick miners are profitable if there is no electrical expense.
Think about what it would cost to buy an RPi, a 13 port USB hub, 10 stick miners, a big battery (L16 HC's), a solar charge controller, a 12Vdc to 5Vdc Buck, and 1-3 solar panels.
There is always more than one way to skin a rat.

I'll get off my soap box . . . . . . .
http://www.ted.com/talks/david_rothkopf_how_fear_drives_american_politics?utm_source=newsletter_daily&utm_campaign=daily&utm_medium=email&utm_content=image__2015-09-14

We left our conversations with Innosilicon at "when you get productions chips we'll buy them in quantity".
To be honest, I'm not the slightest bit interested in buying experimental prototypes which will likely differ considerably from production units.
And having been badly burned in pre-sales schemes before, the statement "not only NO, but F#$% NO" comes to mind.

Acquiring experimental prototype chips (assuming they'd sell them) and forwarding them to GS for board design seems like a huge waste of their time (and a waste of our money).
Given the high probability that the specs on the production chip will be radically different as a result of prototyping.

Put bluntly, if we were going to buy a large quantity of chips (250K-500K) we'd fund the design effort to make one of our own.
This concept is not "off the table".

We have little to no interest in the SFARDS chip, for a plethora of reasons already stated in this thread.

Negative.

BTW, still no response from BM about chips.

I guess the squeaky wheel gets ignored.

Just call me the Minnesota Mushroom: Kept in the cold dark and not feed shit.

One of the things that has always pissed me off about buying electricity is one never knows "how deep" one is in till the bill comes in.
Doesn't give me any current usage feedback and consequently the ability to trim usage to stay within a budget.

First 5 years I lived in my house I made all my electricity with no grid tie backup.
And did NOT live like "Grizzly Adams" (Big screen TV, Microwave, Surround sound stereo, forced air heat, etc. . .)
Was very apparent when consumption rose above norms.
Electricity cost here has tripled since Dec 2005 (the date I ran in grid power) and am re-visiting generating all my own power again given the current trends.

Pretty sad statement when the end consumer can generate electricity cheaper than multi-billion dollar companies.

I beg to differ. I'm in North Central MN (near Leech Lake), > 300 miles north of the IA/MN border.
Our winters are significantly colder and longer than say Des Moines.
About 3 months (June, July, and August) outta' the year ambient temps rise above ~75F


Yup, lotta nukes (Excel Energy).
GRE (Great River Energy) is coal, coal, and coal.
Minnesota Power is a mix of nukes, coal, wind, and up here, wood gasification (this is logging country).

We also have some solar and wind on site.
The $0.0465/KwH is off peak (12 hours/day) which will feed/recharge our UPS's/battery bank at night.
During the day solar augments the $0.12/KwH service.
We are expanding our little solar farm as money permits.
With the goal to be exclusively solar ASAP.
The concept is to buy the power once (a depreciating capex expense) versus suffer an ever increasing re-occurring expense (CODB, Cost Of Doing Business).
Once the off peak is online our aggregate power cost should be around $0.05 - $0.06 / KwH and decreasing as we bring more solar online.

MN was one of the few states that implemented (back in the 70's) fair net excess generation statutes.
Put simply, we can buy power @ $0.12/Kwh and can sell power @ $0.105 /Kwh.
Obviously, if our power generation is in a deficit we want to use everything we generate (grid tieing our generation facilitates this).
IF/when BTC mining becomes unprofitable (which it is becoming) we just turn off the miners and the excess generation becomes the profit center.

Another option is to power the (inefficient) miners exclusively with off peak power and only mine 12 hours/day.

Like so many things in this world, there is no single solution to a problem.
It takes a mix of technologies, intelligently arranged, to solve a problem.
In this case the electrical cost to mine BTC.
I like having options.

To date have not heard back from BM about chips.
That's either real good or real bad.

No secrets here, am working through getting $0.0465/Kwh power online.
Which requires a BUTTLOAD of BS paperwork and all new equipment (Load centers, outlets, conduit, meter bases, etc, ad nausem), which ofc has to be to NEC.

Yet another reason I've never been an Altcoin fan.

You're spot on, the DH key generation is the core hog.
Using large primes as seeds, to some extent, speeds up large DH key generation but also makes the keys more deterministic due to the finite number of seeds (large primes).

To those ends, we are developing, in H/W, a non-deterministic large TRNG (True Random Number Generator) ASIC.
Where large is defined as 16K+ bits, implemented as a USB dongle, and can generate 5-10 TRN's per second.

Most, if not all of the H/W TRNG's I've evaluated are deterministic. As most use the noise off of an electronic circuit as the seed. Given a large enough sample it becomes repetitive (shades of chaos theory), i.e. crackable with a brute force attack. And a  large number of those also have "backdoors", i.e. one can force the TRNG output to a known value through simple code implemented in the corresponding driver software.

Wonder why that is? Ahhhh, the illusion of security/privacy.

So if someone tells me they have a TRNG, IMHO, they're lying through their teeth.
We aim to change that.

Have never been a Altcoin kinda' guy so never looked into it, think that's about to change.
Thanks for the suggestion.
Most of the nodes are nVidia card enhanced (C20's and Quadro 5600's).
My preliminary guess is the power cost will be larger than the revenue stream.
As each node "wastes" a lot of energy on "house keeping" chores.
i.e. power not used in computation.

I'll run some calcs and see if it's worth the hate, hassle, and discontent to run a couple of day test.
And report back with findings.

Sent the 0.242536 BTC ($56) for 2 Compacs to 1BURGERAXHH6Yi6LRybRJK7ybEm5m5HwTr.
Shipping address in message.

Thanks!

Something else, an in-house written DH Key generation routine which uses as it's seed a large Mersenne Prime.
These large primes are no where near as large as the primes GIMPS is working on, but still fall into the definition of large.

Currently, we would like to but are not participating in GIMPS.
Primarily because the cluster sucks a BUTTLOAD of power to run (~50KW) when the nodes are loaded to ~100%.
Consequently, we only "turn it on" when we have jobs for it.

A 16K (actually 16,384) key can take up to 21 days to generate.
So @ 50KW * 24 = 1,200 KWH ($144 @ $0.12/KwH), 1,200 KwH * 21 = 25,200 KwH ($3,024 @ $0.12).

Participating in GIMPS for 30 days (one month) would cost us $4320.
But to be blunt, we just don't have that kind of spare $$$ laying around looking for a home.
As a footnote, the cost and time of generating these rather large DH keys is typically why most VPN providers stick with 2048 and 4196 bit keys (or smaller in some cases).
We have in the past sold computing power for rendering, oil & gas prospecting, and other cryptographic work @ $15-$20 per hour.
Call me a geek but it's really cool to see a couple of hundred cores loaded to ~100% and the power meter spinning like a top.

Have owned and operated a variety of businesses during my 60 years on the planet.
I've found that if a business is operated with a good moral/ethical compass it flourishes.
Those that don't, don't.

BTW, emailed BM about initially purchasing 1K of BM1385's.
Now waiting in "hot standby" for their response.
Would REALLY like to "pull the trigger" on this ASAP.

So, BM if your reading this let's get this vehicle shifted into high gear and get this done.