Like I said, wouldnt count on Ken using the correct terminology .

But there could be some truth to this apparently nonsensical statement; FPGA's do run microcode ('software'). 'our chips' would then have to be referring to FPGA's still. That may sound disastrous that at that point they would still be working on the FPGA prototype, but do keep in mind eASIC nextreme works differently than traditional asics. It uses standard partially processed wafers on which just one layer is changed, and this can be done using an e-Beam; that would be done in house and could be done very fast (within hours literally I believe, but lets call it days).

So here is one hypothesis:
- Nextreme 3 preprocessed wafers are not ready yet or are having issues and they need a respin. Ken may not be allowed to say this publicly.
- meanwhile work is continuing on the ActM FPGA code, changes are made to implement whatever "intellihash" means

The bringing together of those 2 can be done very much faster than with traditional asics, at least for low volume production; its pretty much impossible to guess their timeline, but once working nextreme 3 wafers get delivered, it could be only a matter of days to implement the FPGA code and get working ActM chips (well, wafers) based on them.


I dont quite see how anyone could have thought a startup company producing bitcoin specific hardware could be risk free. FWIW, KnC is also not telling its Neptune customers its using a brand new, barely tested 20nm process, the equipment for which is still being ordered by TSMC. That may or may not pan out, but TBH, I would rate that higher risk than I would have rated eASIC's ability to bring a well understood structured asic migration process to a well understood mature silicon production process. If I were Ken, I would probably have made that same bet.

Like I said, I dont feel too comfortable defending Ken. There is no excuse for stuff like that. That doesnt mean eASIC doesnt bear any responsibility.


Id like to see the exact wording of the original statement if you have it at hand. You can do a respin without having any chips, a respin just means you are redesigning one or more metal interconnect layers (not that I would count on Ken to use the correct terminology).  Did he really have chips and if so, did they even work? Who knows, but I think its highly doubtful. Fact is, there is no record of any 28nm nextreme 3 chip working, easic doesnt even promote it or solicit projects for them. IF the process is healthy, how come?


eASIC would have been suicidal to give iron clad guarantees with large late fees for an unreleased process. More likely is that they cut ActM a sweet deal to be a guinea pig for their new process. A bitcoin asic is pretty much a perfect test chip for a new process, so this could have been a win-win.

I dont feel too comfortable defending Ken here, but fact is eASIC made a public announcement that it would produce these chips using their 28nm Nextreme 3 back in September. A month earlier there was announcement they would use the process for some chip for seagate:
http://www.theregister.co.uk/2013/08/06/seagates_asic_amour/

5 months later, that process is still not commercially available and no word of it on their website. I dont see how you could hold Ken responsible for that. ANd it does to some extend help explain why Ken is so restricted by NDA's.

Ouch. Where those all batch 1 jupiters? Also curious how he is rating KnC's customer support in dealing with the issues.
Anyway, downgraded KnC from 4.5 to 4, as Bobsag isnt exactly the only one reporting fairly severe QC issues.

Aspy?

FWIW, his linkedin profile doesnt suggest he knows the first thing about chip design:
www.linkedin.com/pub/ken-slaughter/22/962/351



I can only guess, but like I mentioned above, ActM would have been eASIC's first project on 28nm with a new, unreleased process called Nextreme 3. Since eASIC are still not promoting it or even acknowledging its existence, Im guessing the problem is right there. Doesnt mean eASIC are clueless, shit happens, particularly if you make your business dependent on new, untested, unreleased processes.  (for the record; Im not accusing Ken of doing something stupid here, in fact a 28nm structured asic produced by a reputable company seemed like a smart bet to me as well at the time)

Thats not how it works in semiconductors. eASIC isnt a fab (though even fabs wont produce your chips without the design passing several validation tests aka tape out), a company like eASIC is contracted to produce working asics.  The whole point of eASIC is a seamless migration from working FPGA to asic conversion. Thats their business. Unless you believe no one ever bothered to test the FPGA, any issues with the structured asic would be on eASIC.

Now Im not 100% sure who wrote the FPGA HDL, but Im fairly sure Ken couldnt write a line of RTL/HDL code to save his life. Considering eASIC offers SHA2 IP cores, its a safe bet they did the entire design.

One last point; easic is still not offering their 28nm nextreme-3 process on their website. If you search easic's website for 28nm you will find precious little besides the press release of their deal with VMC. So its not much of a stretch to think its not working properly yet.

Sounds like at least someone might be delivering early :p

It cant go up by more then 300% at a time IIRC.

Not if they think they can get away with it. For the record, there is already very compelling evidence ghash.io was abused to pull off double spend attacks:
https://bitcointalk.org/index.php?topic=327767.0

Now you may not have sympathy for the victim here,  being a gambling site accepting zero confirmation transactions, but if this is not reason enough for you to switch pools, honestly you deserve what it could lead to.

Are you really wiling to bet the value of of your (and mine and everyone else's) bitcoins to the intention of any single anonymous pool op ?

They used to sell on bitmit, but right now, I dont think they are available anywhere.

Ok, I was not aware of the context of that statement. The statement itself can still be interpreted in several ways, to continue my analogy, HF could agree to send you another paper cheque so are not forced to get a refund through ACH, paypal or even bitcoin for that matter, but refund using the same manner you paid; however given the question he was answering to, thats difficult to argue now.

Until a judge rules on this, we can only speculate. I do see both sides of the argument,  but frankly, I do think HF have a case here. AFAIK, their products were priced in dollar, not bitcoins. Bitcoins were used to fulfill the dollar amount but the number of bitcoins needed to fulfill that amount would have changed daily. Did you think that was unfair too? Its clear to me this wasnt a barter trade.

Another way to look at it; if you had paid with a bank cheque; would you expect or demand to get that very same paper cheque back? Or just the equivalent dollar amount since that cheque was just a way to supply HF with the dollars it demanded for their product? Maybe that bank cheque was signed by a bank director that became a presidential candidate or whatever, and has become worth much more in the mean time, but thats not HF's problem, or is it?

Well, I didnt forget them, but they dont seem to selling (directly) to end users anymore in so far they are still selling anything at all, so Im not sure its that relevant for this thread.

I did add VMC, I somehow seemed to have forgotten them, and Im surprised no one thought of pointing that out.

Ive got a euro card and I pay/withdraw in euro, so no charges for me. Well I believe there is a 2.5 euro fee for every cash withdrawal.

Is anything showing in your bitcurex transaction history?

Do you know anything about operating temperatures?

I would also be interested in a closeup of the original label on the PSU if its still present.

As for efficiency; you may be surprised but most PSU's are less efficient when they are only lightly loaded. So the higher the capacity of your PSU, for a given power draw, the less efficient it will likely be. To be fair, its not easy to find <300W ATX PSU's and some models do well even at very light loads, but unless AMT found one that is > 100% efficient, I highly doubt they will achieve their claimed 80W at the wall. Most reports Ive seen suggest each of the two Technobit boards will consume over 40W by themselves if hashing over 40GH. Add to that the fans, USB hub, Raspberry controller (?) and PSU inefficiency, and well, you do tend to end up with more than 2x40W .

Few people will really care about this in a "80"W miner, but it may provide more insight in how honest AMT is specing their gear.

How can it not be automated if my transactions consistently show up within seconds of getting confirmed, any time of the day?

Please post a screenshot of your account page with the btc deposit address (feel free to blank/cut out all the rest) and a link to your transaction id. If I were to guess Id say you didnt pay transaction fees and your transaction has been stuck in limbo. If so there is nothing Bitcurex can do about that.

Bitcurex are at times slow to process euro withdrawals (slow meaning it could take a day or two, never seen more than 3, typically its the same day), Ive read reports of euro deposits taking some time, but btc transactions, both deposits and withdrawals are as fast as bitcoin allows.