Inaba's request

[Syke]

As a side note, can you point out where I have manipulated the truth?

I'd be more than happy to answer your question, just not in Tom's thread.

No, but BFL as an entity is limiting it's contact with Bitcointalk.  I'm speaking merely as myself and as a pool operator.  On the rare occasions BFL intends to interact with this community, I will post as BFL_Josh. 

You interact on behalf of BFL every day while posting as "Inaba". Take for instance the following lie:

Email office@butterflylabs.com and we will happily refund your order at any time.

"We" obvously means you are speaking for BFL. Since the website clearly states "Payments made for pre-orders of ASIC based products now under development should be considered non-refundable", "at any time" is false because you can refuse to refund an order for any reason you like.

Now this one's a whopper:

But heck, I'm a generous guy, lets cut that in half just for giggles!  405w for 54 GH/s.  Welcome back GPU's!

405 w @ 54 GH/s = GPUs? You're off by 2 orders of magnitude!

I did not decline the bet.

Is the bet on? Nope? You declined the bet.

Power usage is *everything* when it comes to ASIC.

No, actually, when you start mining and how much you pay for hardware are crucial elements in profitability.

I know, for my part, I would not want a device that runs at 200w at 60 GH/s - it would never see a positive ROI.  Even at 150w, your ROI would be measured in years, not months.

Since you don't know what the exchange rate will be, or what the difficulty will be, you cannot state it would never see positive ROI. Neither can you state at 150w it would take years.

Let me just end with one more:

I'm not trolling, I'm a customer asking a legitimate question.

That's just the last few pages of your posts.

[1714937598]

1714937598

[1714937598]

1714937598

[1714937598]

1714937598

[abeaulieu]

excessive compilation of quotes. 'nuff said.

[abracadabra]

:popcorn:

[regular]

This forum is for discussion of hardware.  I think we've given Inaba way too much attention already.  Let's focus on the mining devices instead.

[smoothie]

As a side note, can you point out where I have manipulated the truth?

I'd be more than happy to answer your question, just not in Tom's thread.

No, but BFL as an entity is limiting it's contact with Bitcointalk.  I'm speaking merely as myself and as a pool operator.  On the rare occasions BFL intends to interact with this community, I will post as BFL_Josh. 

You interact on behalf of BFL every day while posting as "Inaba". Take for instance the following lie:

Email office@butterflylabs.com and we will happily refund your order at any time.

"We" obvously means you are speaking for BFL. Since the website clearly states "Payments made for pre-orders of ASIC based products now under development should be considered non-refundable", "at any time" is false because you can refuse to refund an order for any reason you like.

Now this one's a whopper:

But heck, I'm a generous guy, lets cut that in half just for giggles!  405w for 54 GH/s.  Welcome back GPU's!

405 w @ 54 GH/s = GPUs? You're off by 2 orders of magnitude!

I did not decline the bet.

Is the bet on? Nope? You declined the bet.

Power usage is *everything* when it comes to ASIC.

No, actually, when you start mining and how much you pay for hardware are crucial elements in profitability.

I know, for my part, I would not want a device that runs at 200w at 60 GH/s - it would never see a positive ROI.  Even at 150w, your ROI would be measured in years, not months.

Since you don't know what the exchange rate will be, or what the difficulty will be, you cannot state it would never see positive ROI. Neither can you state at 150w it would take years.

Let me just end with one more:

I'm not trolling, I'm a customer asking a legitimate question.

That's just the last few pages of your posts.

OWNAGE

[cheebydi]

This forum is for discussion of hardware.  I think we've given Inaba way too much attention already.  Let's focus on the mining devices instead.

+1

Everyone who is able to read and think clearly should've made up his mind by now.

[mc_lovin]

Bit confusing, but the answers he gives are not really making a lot of sense considering.  If an ASIC = 150x a GPU's [mining] power, welcome back GPUs, huh?

[Keefe]

You should be looking at W/$. Capital investment in bitcoin mining hardware will likely be similar in the ASIC era as in the GPU/FPGA era, adjusting by block reward and exchange rate of course. So forget the hashrate. $1070 of GPUs use about 800W of power. $1070 of BFL FPGAs use about 150W of power. If a $1070 bASIC uses 405W, we're taking a step backward to almost GPU levels. $1070 of BFL ASICs should use only 50W of power, a significant step forward. I believe Inaba was thinking along this line when he said that.

[cablepair]

You should be looking at W/$. Capital investment in bitcoin mining hardware will likely be similar in the ASIC era as in the GPU/FPGA era, adjusting by block reward and exchange rate of course. So forget the hashrate. $1070 of GPUs use about 800W of power. $1070 of BFL FPGAs use about 150W of power. If a $1070 bASIC uses 405W, we're taking a step backward to almost GPU levels. $1070 of BFL ASICs should use only 50W of power, a significant step forward. I believe Inaba was thinking along this line when he said that.

Please do some research before posting information you know nothing about

For one thing a bASIC will not use anywhere near 405watts, I dont know where this number came from but its complete crap.
But even if it did use 405 watts (which it does not) it is still LIGHT YEARS ahead of GPU mining.

Please refer to the Bitcoin Mining Hardware comparison chart here:


https://en.bitcoin.it/wiki/Mining_hardware_comparison


Show me any possible way you can mine with GPU at 54Gh/s and only use 405 watts?

Are you really that clueless or are you just a sock puppet because I really find it hard to believe you are that clueless.

Even with the most efficient GPU 7970 in the absolute best conditions to achieve 54Gh/s

You would need 68 7970's which would use a total of 14552 watts to hash at 54Gh/s

how does that even come close to comparable?


 Please do not re-post bullshit manipulative lies.

bASIC mining units will be competitively efficient with any of our competitors.

and by the way not one single Bitcoin ASIC manufacturer has posted any PROOF of electrical efficiency.

Stop spreading lies please.

[ralree]

Quote removed because of edit.  See Above.

Very nice response to the trolling.  Nice job!

[cablepair]

darn it, you quoted me before I could edit my post

even though its getting harder and harder to do so I am trying very hard not say anything bad about my competitors, so I re-edited the post to tone it down a bit.

its ok though,,,,

It's just so frustrating, all these lies and manipulating people in the name of greed.

Anyone that gave a shit about this community would not do these things. The truth is the truth and people need to know it.

[Tinua]

In a few months, Your 54GH/ s seems to be the same as 800 MH/s today! Thus, the comparison W/$ on the purchase of an ASIC seems already correct!
So everything is over 150W/1000$ seems for the European market not so great anymore!
Just my impression!

[gmaxwell]

You should be looking at W/$. Capital investment in bitcoin mining hardware will likely be similar in the ASIC era as in the GPU/FPGA era, adjusting by block reward and exchange rate of course. So forget the hashrate. $1070 of GPUs use about 800W of power. $1070 of BFL FPGAs use about 150W of power. If a $1070 bASIC uses 405W, we're taking a step backward to almost GPU levels. $1070 of BFL ASICs should use only 50W of power, a significant step forward. I believe Inaba was thinking along this line when he said that.

[...]
Show me any possible way you can mine with GPU at 54Gh/s and only use 405 watts?

While I don't agree with Keefe, he's actually making a coherent argument here.

First, as you know— the earnings from mining don't depend on your absolute hashrate. 1MH/s, 1GH/s, 1TH/s whatever. What matters is your rate relative to the other miners. Keep this in mind.

If you assume that all mining is done by initial purchase cost limited miners who bought cards just for mining, and e.g. that as many people who spent $1070 on GPUs will spend $1070 on ASICs  ...  then eventually your $1070 will have the earnings equivalent of what $1070 in GPUs has today, but a higher operating cost because it will draw more power than $1070 in GPUs. (well maybe, but thats the argument).

Of course this is all crazy and wrong, because (1) there is some unknown but non zero amount of botnet miners who are going to become insignificant, (2) there is some unknown but non zero amount of casual miners using their primary GPUs who are going to become insignificant, (3) OPERATING COST IS A MAJOR DRIVING FACTOR IN RUNNING A GPU FARM, it's probably best to assume that large miners would match their operating costs, not their initial costs (4) some miners have already spent their money on GPUs and won't continue anymore, (5) some miners felt more comfortable investing in GPUs which have some resale value if bitcoin fails. (6) People holding out on ASICs until they are for sale for immediate delivery or for second-generation products or for the effect of competion to lower prices. etc.  There are also reasons we could expect things to be less profitable: people holding out on GPU purchases in anticipation of asics, horded gpu earnings being released to buy asics, big asic farms should be dollar per dollars much easier to manage (GPU FAILURES SUCK) encouraging bigger ones. (not to mention the subsidy decreases!)

The model is all wrong, but it's not crazy or abusive.

Please don't let the people here stress you out.  The proof is in the pudding. Your product will amaze people. Spend your limited mental cycles on making it the best you can. If all goes well, or if all blows up.. it will be what it will be and none of these arguments matter.

[squid]

You should be looking at W/$. Capital investment in bitcoin mining hardware will likely be similar in the ASIC era as in the GPU/FPGA era, adjusting by block reward and exchange rate of course. So forget the hashrate. $1070 of GPUs use about 800W of power. $1070 of BFL FPGAs use about 150W of power. If a $1070 bASIC uses 405W, we're taking a step backward to almost GPU levels. $1070 of BFL ASICs should use only 50W of power, a significant step forward. I believe Inaba was thinking along this line when he said that.

[...]
Show me any possible way you can mine with GPU at 54Gh/s and only use 405 watts?

While I don't agree with Keefe, he's actually making a coherent argument here.

First, as you know— the earnings from mining don't depend on your absolute hashrate. 1MH/s, 1GH/s, 1TH/s whatever. What matters is your rate relative to the other miners. Keep this in mind.

If you assume that all mining is done by initial purchase cost limited miners who bought cards just for mining, and e.g. that as many people who spent $1070 on GPUs will spend $1070 on ASICs  ...  then eventually your $1070 will have the earnings equivalent of what $1070 in GPUs has today, but a higher operating cost because it will draw more power than $1070 in GPUs. (well maybe, but thats the argument).

Of course this is all crazy and wrong, because (1) there is some unknown but non zero amount of botnet miners who are going to become insignificant, (2) there is some unknown but non zero amount of casual miners using their primary GPUs who are going to become insignificant, (3) OPERATING COST IS A MAJOR DRIVING FACTOR IN RUNNING A GPU FARM, it's probably best to assume that large miners would match their operating costs, not their initial costs (4) some miners have already spent their money on GPUs and won't continue anymore, (5) some miners felt more comfortable investing in GPUs which have some resale value if bitcoin fails. (6) People holding out on ASICs until they are for sale for immediate delivery or for second-generation products or for the effect of competion to lower prices. etc.  There are also reasons we could expect things to be less profitable: people holding out on GPU purchases in anticipation of asics, horded gpu earnings being released to buy asics, big asic farms should be dollar per dollars much easier to manage (GPU FAILURES SUCK) encouraging bigger ones. (not to mention the subsidy decreases!)

The model is all wrong, but it's not crazy or abusive.

Please don't let the people here stress you out.  The proof is in the pudding. Your product will amaze people. Spend your limited mental cycles on making it the best you can. If all goes well, or if all blows up.. it will be what it will be and none of these arguments matter.

I think the annoyance is just over the fact that Josh was using the MMQ FPGA as a "model" for power use of the bASIC. Which is ridiculous. It would be me saying that since all BFL has are simulation, not actual proof of power usage that I don't believe their numbers and will use the FPGA model to predict the power of their ASIC.

This is just number manipulation/nonsensical trolling at its finest.

How much does power difference matter? It will be impossible to say until we see the market shifting.

[nick0016]

Some interesting information posted here. But still this is the hardware forum...

[Bogart]

Some interesting information posted here. But still this is the hardware forum...

I'm just glad to see it put in its own special thread rather strewn through other threads that were really meant for other things.

Thanks for that, OP.

[ralree]

darn it, you quoted me before I could edit my post

Edited to protect the innocent

[Keefe]

Wow, I didn't expect Cablepair to blow up at me of all people. I've posted in the past on Cablepair's side against Inaba's claims that the bASIC won't be competitive. Are you upset about my latest post just because I referenced Inaba's 405W figure? I made sure to say "IF" a bASIC uses 405W. As Gmaxwell said, I am making a logical argument, despite oversimplifying. FWIW, I've ordered 4 of the bASIC 54GH units, so Cablepair currently holds ~$4400 of my money. I feel a little insulted that I get called a troll in return.

[Keefe]

Of course this is all crazy and wrong, because (1) there is some unknown but non zero amount of botnet miners who are going to become insignificant, (2) there is some unknown but non zero amount of casual miners using their primary GPUs who are going to become insignificant, (3) OPERATING COST IS A MAJOR DRIVING FACTOR IN RUNNING A GPU FARM, it's probably best to assume that large miners would match their operating costs, not their initial costs (4) some miners have already spent their money on GPUs and won't continue anymore, (5) some miners felt more comfortable investing in GPUs which have some resale value if bitcoin fails. (6) People holding out on ASICs until they are for sale for immediate delivery or for second-generation products or for the effect of competion to lower prices. etc.  There are also reasons we could expect things to be less profitable: people holding out on GPU purchases in anticipation of asics, horded gpu earnings being released to buy asics, big asic farms should be dollar per dollars much easier to manage (GPU FAILURES SUCK) encouraging bigger ones. (not to mention the subsidy decreases!)

Total network hash power right now is around 23 TH. That's like ~$12M of GPUs, earning ~$2.6M per month (at $12/BTC) minus power cost. When the block reward is halved, $6M of GPUs earning $1.3M per month would have the same ROI. When $1300 buys a 60 GH unit, $6M will buy ~280 TH, and will have a similar ROI. So when looking at power draw of this new hardware, you should account for going from 23 TH to 280 TH. You should adjust a unit's hashrate by 12x to get a fair comparison.

I believe that the predominant factor determining how much hardware is mining is ROI. And I believe that it's decided at the margin, by those with cheap power and lots of capital to work with. Those people can buy and setup additional GPUs and FPGAs as the exchange rate goes up and difficulty lags, until price/difficulty is back to where they are comfortable with the ROI.

The cost of power, especially for FPGAs and ASICs, isn't that significant compared to the hardware cost. If you spend $1000 on a unit, and only $50 on power for a year, and you earn $2600 in a year, power doesn't make a big % impact on ROI.

I don't think botnets and casual miners affect the equation. If there is 2 TH of botnets, that just means the professional miners will likely run 2 TH less hardware, because they're looking at ROI. And if some people decide to stop buying more hardware, as long as there are a few professional miners who will take up the slack it won't matter.

Now I do agree with at least one of Gmaxwell's points: GPUs have value outside mining, which can make a big difference in the risk calculation. So maybe professional miners will want a better ROI to compensate. Also, ASICs will likely drop in price (per GH) alot over the next year, as companies like BFL and BTCFPGA fight for market share with products with huge marginal profit (after NRE). Those two factors could make a big difference in where price/difficulty stabilizes.

[Coinoisseur]

Wow, I didn't expect Cablepair to blow up at me of all people. I've posted in the past on Cablepair's side against Inaba's claims that the bASIC won't be competitive. Are you upset about my latest post just because I referenced Inaba's 405W figure? I made sure to say "IF" a bASIC uses 405W. As Gmaxwell said, I am making a logical argument, despite oversimplifying. FWIW, I've ordered 4 of the bASIC 54GH units, so Cablepair currently holds ~$4400 of my money. I feel a little insulted that I get called a troll in return.

Don't underestimate people's ability to ignore the words "if" "possibly" "perhaps" "supposedly" and such. I think Cablepair has a legitimate worry of his ASIC product being unfairly associated with a made up power number.

Also interesting how a standalone post can go off topic so quickly, I believe this thread was about demonstrating Inaba/BFLJosh's identities are not well firewalled and citing possible spurious statements of his.

On topic: Has BFL provided their operating address yet? Inaba/BFLJosh indicated he would be able to do so "soon" as of September 26:

I'm hoping next week we will begin the process.  I was hoping the end of this week, but it's taking a bit longer than I'd like.