Topic: [PicoStocks] 100TH/s bitcoin mine [100th] - page 88. (Read 469872 times)

I think the PicoStocks project is the main one. A group of people is investing quite big money to make it running.
The 100TH/s bitcoin mine is an auxiliary project, started as a mean to promote PicoStocks.

Just kick people in the balls. It will have the same effect as "investing" in mining operations but be less painful.

With a similar level of optimization as BFL, 32nm would probably do it. That is of course assuming BFL hits its power targets. It's not impossible that they could come in with a more optimized design than BFL has on a 40nm node and meet those specs.

That being said, I haven't seen bitfury post here and there's nothing to suggest that they will produce anything close to that in the timeline they suggest. If it's true I wish them the best, but there's all kinds of sayings about the burden of proof on exceptional claims, and there's nothing here to back that up yet.

I completely agree with MPOE: tytus' proposal seems to be the one of a scammer.

Other data point: he claims "0.2-0.3 Watt per GH/s". This is completely unrealistic. This means 3300-5000 MHash per Joule. To get anywhere close to this, you would need to design the chip around the 22-32nm process node!

Perhaps I read the timeline wrong, but it sounds like they're planning on designing the PCB after the chips are packaged. I'm not suggesting that they need to have the full batch of PCBs made at the same time, but I don't see any reason to not have the preliminary testing PCBs done and get on ordering them early, as well as have the production PCB designed at the time that the chips roll off the line. If you have to tweak the design you can, but waiting seems counterproductive.

Thanks :-)

We will do only one Engineering run first (requirement of the foundry). The production run will follow but this is of course a huge delay.

The reason for not designing the PCB in advance is that:
1) the chips have a build in power management / leveling system that we want to use (to save on power and additional pcb components: to power the boards and a string/series of chips with let's say 7.2V directly)  ... this is a crazy idea and we have to check if this works (we must avoid resonance buildup and power voltage osculations) before ordering many boards [[ comments here are very welcome !!! :-) ]]
2) we want to use the PCB for cooling by placing (low power consuming) chips far apart (no fans and possibly no radiator). Power specs are now from simulation only. If we fail to meet them we have to cover the potential losses for the investors, but we also will have to redesign boards to deal with increased heat production.
3) BitFury can do test PCBs "in house" in 2-3 hours. This will help designing optimal boards before large order.

We will rethink this of course. If we decide to stay with the standard powering solution (adding psu for chips on board) the boards could be designed faster.

Two reasons:

1) PCB change may be a last-resort fix to avoid mask change.
2) PCB for testing/debugging may have significant (but trivial) layout differences than the PCB for mass production.

Remember that there are actually two production runs: (1) short engineering run and (2) normal production run. Engineering run ICs may have different package and/or pinout than the production run ICs.

Disclaimer: I'm not bitfury, nor do I speak for him.

Why wouldn't you have the PCBs designed and ready during the week or two after the wafer starts instead of waiting to the end? You'll leave yourself more time, and save a lot of money if you let them get build over a few weeks instead of in under a week.

Thank you for your reply. Your mention of "trusted account" with an exchange (implied MtGox) made me think about their history of bitomat acquisition together with the bank contacts in Poland.

I know. This is a problem, but there is nothing we can do about it :-(

None
None, but I have to look at that :-)

Deal :-)

Far as I can see, you're here to scam. The reasons are quite simply:

1. You're nobody, I couldn't tell you from Adam. Tytus, what's that. Wot? The cost to make you, entirely, is about fiddy cent. (Sure, you may or may not be related to Leszek Rychlewski, who'd be a moderately cited biology researcher. On the strength of your presentation that connection is not worth investigating.)

2. You came up with what you call an exchange, for no apparent reason. It doesn't exchange anything. The worthless assurances of some Marshall Islands shell "repaying" anyone for anything are of no value and no interest.

3. You came up with a scam for that exchange, specifically, an imaginary set of 1000 boards that Tom was selling you (in spite of bASIC not making that many boards to begin with and you not being anywhere on the lists of people that ordered).

I did point out to you then you're lying through your teeth, and your response was that *I* should check up on your claims. No receipts, no proof of anything. You could be Inaba if we didn't already have an idiot by that name around.

Conveniently enough for you Tom blew up, and so you're roughly in the situation of the various scammers operating out of GLBSE: they get to pretend like it was all Nefario's fault. Oh, yes, I forgot, you lost a million dollars with the bASIC blow-up, how fortunate for us that you're so cavalier about that disaster, picked yourself up and are back for more. Innit easy to get through due diligence for imaginary fortunes and imaginary investments? How strange that Tom never had half that much money to refund in the first place.

4. You came up with a slightly larger piece of bs. Who is bitfury? Oh, wait, nevermind, "BitFury Group facilitating development of BitCoin-related software, hardware and services". Great. They made a splash last summer with their insane "licensing" deal. Meanwhile they draw 7MW per Gh or something like that, and nobody (including the manufacturer) even knows what the units cost. Let alone that there were a grand total of two units built overall, during the entire 10 month lifetime of the project. As a bonus, MegaBigPower.com is some bs site registered two weeks ago, in no position to "operate" anything whatsoever. To make it perfectly clear: 100 TH made out of bitfuries is much like an aircraft carrier made out of old refurbished Ford Model Ts: not seaworthy.

5. If you think that is a business plan you have absolutely never either seen one or done business.

I get that you think Bitcoiners are idiots, and most are. The important point is that those who are have no BTC, and those who still have BTC aren't idiots. Your deluded notions that anyone will actually put as much as fifty dollars into this patent nonsense is about as amusing as the idiocy coming from the bitfinex corner. I get that you think you can buddy-buddy and "act professional" and it'll wash like it has in the past. You're wrong. It won't.

To quote my ever-watchful boss,

Bitfury has already been running simulations.  He's in Kiev.  Les is in Poland.

From the biz plan - the July date is appropriately conservative:
"Chip design will be finalized on 2013-02-15. The wafer fabrication will start 2013-03-01.
Wafers will be completed on 2013-04-20. Small subset of chips will be packaged and
tested and PCBs will be designed by 2013-04-30. The remaining chips will be packaged
and  PCBs  will  be  assembled  by  2013-05-20.  Hardware  will  be  assembled  at
MegaBigPower.com on 2013-05-30. Mining could start on 2013-06-01, but we expect
unforeseen delays of approximately 1 month. Thus mining will start at the datacenter on
or before 2013-07-01."

Well ... since I've already had a chat with the 'other' person about this, I though it worth mentioning it here.

Dude, bitfury is the lead designer at their shop. If I know anything about the people of bitfuty's calibre I'll say that he will have driver's core already debugged on the simulator before the tapeout. This really isn't looking like another seat-of-the-pants outfit.

Bitfury had been working on the Bitcoin ASIC since early last year, and on the processes available through Europractice since August 17 of last year. I though he was Russian, Ukrainian or Bielorussian and I pointed it to him that he has that option available:

https://bitcointalksearch.org/topic/m.1105513

I didn't know that he's Polish or had choosen to work through a Polish scientific establishment. Through Europractice he should have no problem accessing the latest 40nm and 28nm processes at deep discount.

I really like what you're doing, and I'm very interested, but this is kind of a dealbreaker for me. And for a lot of people, I'm sure.

A) What is your relation with former bitomat.pl and its former operator Bartlomiej Szabat?

Edit:

B) What is your relation (if any) with Enterpoint and its recently announced 100THash Goliath project?

https://bitcointalksearch.org/topic/goliath-130386

Thanks.

Waiting for a dev board so I can write the cgminer driver for you

Edit: of course, contact me if you want any suggestions about the MCU design (not doing the design, just optimal details of it's design)