Topic: [BTC-TC] BTC-Mining - page 6. (Read 22665 times)

When I said efficiency, I meant power efficiency as in GH/W.

The price/GH you are talking about is something else entirely. Development of an ASIC is very expensive, production is almost free. As a result,  BFL can price those things almost arbitrarily. I made the estimates in a different thread, assuming BFL used an old 130nm process, a single wafer would yield somewhere between 1 and 10 TH worth of asics. Such a wafer costs about $1000 to process (and a few dollars per chip on top of that for cutting and packaging).  IOW, the PCB, assembly, housing and power supply will cost a lot more than the chip itself. That gives them enormous pricing flexibility.

So its not because BFL now charges "only" 10x less per GH than for FPGAs, that this difference will not go up substantially when difficulty goes up and market value of these ASICs comes crashing down proportionally. Unlike GPUs or FPGAs, these asics can not be sold for anything other than bitcoin, so their price will follow bitcoins price/difficulty from the current peek, to a bottom thats somewhere between one and two orders of magnitude lower. So yeah, Im predicting 1 TH for less than $3000 over the next year, maybe two, depending how many people buy in to this crazy gamble. Or less if a competitor emerges.

Smart move. 

Not sure for Jalapeno vs Single or SC Single vs Single, but I calculated for MiniRigs only and the efficiency difference was 20x
MiniRig: 25 000 mhash/15 000 USD = 1.66 mhash/USD
SC MiniRig: 1 000 000 mhash / 30 000 USD = 33.33 mhash/USD

It's not to say that they won't dislodge FPGAs. We're going to motion for going with ASICs instead of FPGAs.

However, from what I read about FPGA vs ASICs, it seemed like performance would be closer to a range of 2 to 4 the performance of FPGAs. 20x just seemed wrong. Although I do have to admit I do now know about specifics for SHA256 ASICs. I do know they should be a lot more power efficient. Around 10x the efficiency. It's really the expected extra performance I had in mind (2 to 4). 20x seems unrealistic, even for ASICs. ALTHOUGH I've heard they did use older chips for those FPGAs. That might have played in there too, although a 20x efficiency jump really is a far shot from what I estimated from reading about the difference between FPGAs and ASICs.

You are so wrong. Of course I cant vouch for BFLs promises (particularly when it comes to delivery times), but anyone not believing the numbers BFL publishes are even possible with an ASIC, simply doesnt know what he is talking about. Go ahead and request a free asic here:

http://rijndael.ece.vt.edu/sha3/sha3chip

Or at least read the datasheet. Yes, its a miniature SHA3 test chip but it also has a tiny SHA2 engine as reference. You will find that despite being build on an old 130nm process, despite not being optimized for bitcoins peculiar way of using SHA, despite not even being optimized for generic SHA2 hashing (big compromises were made to include all the SHA3 test algorithms), its actual performance/W seems in line with BFLs claims and some 50x better than current FGPA solutions. 

Bashing BFL is popular around here, and they certainly deserve some flack,  but stop putting your head in the sand when it comes to ASICs. They do have the potential of achieving two orders of magnitude higher performance and higher efficiency than FPGAs, while their variable costs are negligible. Once they start shipping,  28nm FPGAs  will be as relevant for mining as 32nm CPUs were last year.

Mining returns of 28.83860097 BTC paid @ 0.01491909 BTC per share for the week of 17th June to the 23th June of 2012

If you look a few posts above you I was asking about that thing to. The answer was let's wait for the actual specs cause most people don't believe they'll be that high.

On they other hand don't forget that 40% of the mined bitcoins are reinvested. So if things go as intended sometime we will probably buy ASIC hardware too.

Also keep in mind that BFL promises (well, for what their promise is worth) that they will buy back FPGA miners at full cost when you buy ASIC from them. I have totally no idea how they are going to do that to be honest though.

This is actually great. It means you are in a good position for buying ASICS. Change thread title to reflect this!

Exactly as Tritonio said. They don't gain or lose any value.

The same as the one that we got them at. For every 1BTC we gave him we will get the same 1BTC back.

At what price will amazingrado buy back the bonds?

As Diablo said, their announcements have always been a bit... blown into myth material. ASICs shouldn't end up being much more efficient than FPGA. They cost more to design but end up being cheaper to produce so it is good for volume. But they don't end up being a whole lot more performant, even if they are. Not on such a large magnitude. I have high doubts they could pack that kind of processing power in less space than the current FPGAs they use. Probably just an announcement to get some talk about them and visibility. I don't say I approve such tactics but it seems to have worked for them. They end up with acceptable specs, lowest market price. So even if they don't reach their promises, people still buy their equipment. For now I'm waiting for the actual performance expected and product to be put on market to further consider the new ASICs.

No one trusts their specs or delivery dates. 28nm also may completely crush it.

Aren't we buying their minirig though? Or you mean nobody trusts their specs or their delivery dates?

Don't bother bringing that up. No one trusts BFL.

http://news.yahoo.com/butterfly-labs-announces-next-generation-asic-lineup-054626776.html

Any ideas when are those coming out?

If we get the minirig, can we sell it back for $15000 as this article says? (or so I understand)

Maybe we could spend the first 15K we gather in the new single units? We could get more than 10 of them for a total of 400Ghash/s.

And then when we collect the next 30K we can get the new minirig for another 1Thash/s.


I still can't believe those numbers. I am reading them again and again... Is this going to happen during 2012???

EDIT: I just read on their site that they will deliver probably on October.

ok, sounds good.

Almost, as long as Amazingrando doesn't fail to purchase the bonds back. He's been pretty reliable and trustworthy so far, investing in his own mining operation and giving out dividends on his Bitbond title and to us. He does have a lot of experience in bitcoin mining too. I'm not really worried about that, although something catastrophic could always happen to him. Late payments/default risk are limited to the issuer not being able to continue operating his mining operation and so far it seems more than a healthy one. So pretty much the exact same risks you'd expect from placing your funds with any mining title on GLBSE or with BTC-Mining.

i have another question:

is it 100% guranteed, that the IPO bitcoins which are currently bound to the mining-contract will be returned at the end of august?!

or is theire any risk, you hide?

Mining returns of 30.1346067 BTC paid @ 0.0156057 BTC per share for the week of 10th June to 17th June of 2012

Also will update the main post about the new ASICs announcement.

I own 658 of BTC-Mining atm.