1GH/s, 20w, $700 (was $500) — Butterflylabs, is it for real? (Part 2)

[Raoul Duke]

I am borderline retarded when it comes to computers.

That's because your fatness doesn't let you get closer to the computers...

[1713888949]

1713888949

[RandyFolds]

I am borderline retarded when it comes to computers.

That's because your fatness doesn't let you get closer to the computers...

Which is why I rest a trackball mouse and keyboard on my belly.

[fred0]

You guys didn't give BFL a chance.  From day one on the original thread you were accusing them of being scammers.

Because we didn't believe their claims. And we were right. There is no 1GH/s, 20w, $700, miner.

after taking 2 days to read all the threads....  give em a break that was before the POC.  now it is 800+ at 80w. still awesome! with a good chance at 60-70W with the better psu

No, POC is still 4-6 weeks from delivery...just like three months ago. I could get you some (READ: ONLY TWO UNITS HAVE BEEN PROVEN TO EXIST) friggin' FPGAs built if you gave me thousands of dollars and a few months, and I am borderline retarded when it comes to computers.

How many need to exist? My understanding is that first pre-orders should be delivered in mid-January.  I would rather have a late widget than no widget, or a widget that is in imminent danger of failure due to insufficient power handling/capacity.

Isn't the purpose of the prototype to shake out these types of design deficiencies?  I don't think it would make sense to fab 100 of these and then have 100 fail in customers' possession.  Sure the development process has revealed mistakes, but that is one of the characteristics of development.   

[RandyFolds]

You guys didn't give BFL a chance.  From day one on the original thread you were accusing them of being scammers.

Because we didn't believe their claims. And we were right. There is no 1GH/s, 20w, $700, miner.

after taking 2 days to read all the threads....  give em a break that was before the POC.  now it is 800+ at 80w. still awesome! with a good chance at 60-70W with the better psu

No, POC is still 4-6 weeks from delivery...just like three months ago. I could get you some (READ: ONLY TWO UNITS HAVE BEEN PROVEN TO EXIST) friggin' FPGAs built if you gave me thousands of dollars and a few months, and I am borderline retarded when it comes to computers.

How many need to exist? My understanding is that first pre-orders should be delivered in mid-January.  I would rather have a late widget than no widget, or a widget that is in imminent danger of failure due to insufficient power handling/capacity.

Isn't the purpose of the prototype to shake out these types of design deficiencies?  I don't think it would make sense to fab 100 of these and then have 100 fail in customers' possession.  Sure the development process has revealed mistakes, but that is one of the characteristics of development by marketing amateurs.  

ftfy

How many iphone releases have been delayed because they didn't work?

And as far as how many should exist? At least one development board to serve as POC and measure exact speeds and power consumption, the only really relevant numbers in their whole game. That obviously didn't happen until pretty recently.

[fred0]

How many iphone releases have been delayed because they didn't work?

http://www.cnn.com/2011/11/10/tech/mobile/iphone-4s-battery-fix/

I don't think the Apple releases were delayed, but maybe they should have.

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

[DeathAndTaxes]

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

I doubt it.

The rig box "magically" has twice the power efficiency of the single (and despite being asked in this thread) BFL has provided no explanation.  I asked "is that power efficiency based on a working product or only on a simulation" without a response.

[fred0]

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

I doubt it.

The rig box "magically" has twice the power efficiency of the single (and despite being asked in this thread) BFL has provided no explanation.  I asked "is that power efficiency based on a working product or only on a simulation" without a response.

2500 / 32 = 78.125 W/upgrade slot.  This is consistent with the power reported for the POC single. The modules are not the same as the singles, and the rig box is supposed to have hardware unique to itself.

And I'm weary of this BFL bashing.  If you think there is something fishy going on, then by all means do not spend any money with BFL. Not revealing their IP does not mean they are scammers, only capitalists.

Let's see where we are in mid-January, better data, better conclusions.

[DeathAndTaxes]

2500 / 32 = 78.125 W/upgrade slot.  This is consistent with the power reported for the POC single. The modules are not the same as the singles, and the rig box is supposed to have hardware unique to itself.

I don't think you understand what power efficiency means.

Single:
800MH/80W = 10MH/W

RigBox:
50.4GH / 2500W = 20MH/W

For a company who's prior power estimate turned out to be based only on ISE simulation (and was off by 400%) I don't think calling a request for clarification on a 100% power efficiency improvement to be "bashing". 

And I'm weary of this BFL bashing.  If you think there is something fishy going on, then by all means do not spend any money with BFL. Not revealing their IP does not mean they are scammers, only capitalists.

An answer:
"yes this is based on a physical functioning product and we have verified it will get 50.4GH on 2500W or better" or "no this is based only on simulated data, the same flawed simulation which indicated 19.8W for the single" is hardly revealing IP.  It is a pretty basic question.

[RandyFolds]

Let's see where we are in mid-January, better data, better conclusions.

That's basically what I said in October...

If you grow weary of BFL bashing, perhaps you should stop reading this thread. They haven't really given anyone a reason to stop bashing them. Most of the wild allegations have been proven true.

Where exactly is that ASIC again?

[plastic.elastic]

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

I doubt it.

The rig box "magically" has twice the power efficiency of the single (and despite being asked in this thread) BFL has provided no explanation.  I asked "is that power efficiency based on a working product or only on a simulation" without a response.

2500 / 32 = 78.125 W/upgrade slot.  This is consistent with the power reported for the POC single. The modules are not the same as the singles, and the rig box is supposed to have hardware unique to itself.

And I'm weary of this BFL bashing.  If you think there is something fishy going on, then by all means do not spend any money with BFL. Not revealing their IP does not mean they are scammers, only capitalists.

Let's see where we are in mid-January, better data, better conclusions.

Wow, how does that show efficiency genius?

Why i'm not surprised knowing you're a sheep listening to BFL's bullshiet.

[Raoul Duke]

I just gotta leave this here: https://bitcointalk.org/index.php?topic=52081.msg663353#msg663353

How come a guy who doubted 1GHs/20W was possible is listening(and believing) to rumours about 1GHs/10W??

Funny, indeed...

[fred0]

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

I doubt it.

The rig box "magically" has twice the power efficiency of the single (and despite being asked in this thread) BFL has provided no explanation.  I asked "is that power efficiency based on a working product or only on a simulation" without a response.

2500 / 32 = 78.125 W/upgrade slot.  This is consistent with the power reported for the POC single. The modules are not the same as the singles, and the rig box is supposed to have hardware unique to itself.

And I'm weary of this BFL bashing.  If you think there is something fishy going on, then by all means do not spend any money with BFL. Not revealing their IP does not mean they are scammers, only capitalists.

Let's see where we are in mid-January, better data, better conclusions.

Wow, how does that show efficiency genius?

Why i'm not surprised knowing you're a sheep listening to BFL's bullshiet.

I said nothing about efficiency,  read much?  Welcome to ignore.

[bulanula]

I just gotta leave this here: https://bitcointalk.org/index.php?topic=52081.msg663353#msg663353

How come a guy who doubted 1GHs/20W was possible is listening(and believing) to rumours about 1GHs/10W??

Funny, indeed...

Just a rumor that will probably never become true.

[plastic.elastic]

In any case, BFL seem to be nearing the finish line, let's hope there aren't more surprises(bad surprises).  Of course, this could be repeated with the Rig Box and "Super Computer", but I think that the school of hard knocks has taught a lesson.

I doubt it.

The rig box "magically" has twice the power efficiency of the single (and despite being asked in this thread) BFL has provided no explanation.  I asked "is that power efficiency based on a working product or only on a simulation" without a response.

2500 / 32 = 78.125 W/upgrade slot.  This is consistent with the power reported for the POC single. The modules are not the same as the singles, and the rig box is supposed to have hardware unique to itself.

And I'm weary of this BFL bashing.  If you think there is something fishy going on, then by all means do not spend any money with BFL. Not revealing their IP does not mean they are scammers, only capitalists.

Let's see where we are in mid-January, better data, better conclusions.

Wow, how does that show efficiency genius?

Why i'm not surprised knowing you're a sheep listening to BFL's bullshiet.

I said nothing about efficiency,  read much?  Welcome to ignore.

What an idiot, you replied to a post with POWER EFFICIENCY in question.

Ignore my ass, you know you will read this.

[makomk]

I just gotta leave this here: https://bitcointalk.org/index.php?topic=52081.msg663353#msg663353

How come a guy who doubted 1GHs/20W was possible is listening(and believing) to rumours about 1GHs/10W??

Funny, indeed...

The key part there is the "ETA 6 months". The problem with BFL's original claims wasn't that 1GHs at 20W is impossible, but that it made no sense for a company to suddenly appear selling boxes that achieved it because it requires ASICs with long lead times and huge up-front costs. If you're going to be producing this kind of ASIC, it makes sense to announce it up-front in order to discourage anyone else from trying to manufacture them too and cutting into your profit margins.

[Epoch]

Where exactly is that ASIC again?

Do you want to know? I'll be glad to tell you: [Tycho] AKA Deepbit is developing it. You're welcome.

Don't tell me: another 1.05GHash/s, 20W, $499 preorder miracle box? Seems we're coming full circle.

Realistically, further FPGA/ASIC development is inevitable as long as bitcoin mining remains profitable. We can look forward to another 12 months of 50BTC/block reward, but after that the reward is cut in half: 25BTC/block. Few expect the BTC->USD exchange rate to be affected by this much, which will leave a lot of the existing GPU crowd in unprofitable territory if BTC prices are in the $2-$4 range by then. FPGA and ASIC miners are the only real option, but whether they are viable depends on their ROI.

Few will be willing to jump on the FPGA/ASIC train if the ROI isn't attractive. The resale opportunities for an FPGA/ASIC miner are extremely limited since they are very much a niche product; whereas GPUs can be resold at attractive prices without issues.

BFL's 'product' (a loose term, granted, until they actually ship) has a current break-even time of 10 months. Similar to a 5870 or other efficient GPU. What could you resell the BFL box for after 10 months, with no warranty? Half price? I can buy 2 5870s for $400 today and be confident that I can resell them for 10% less ($360) in 10 months.

Even if a newer FPGA or ASIC design achieves 1/10th the power draw of the BFL product at similar hashrates and cost, it doesn't have much advantage from an ROI perspective: the payback for such a device (even assuming 0 power draw) would be ~9 months compared to BFL's ~10 months. BFL's product will be available (theoretically) in month. Any other vaporware FPGA/ASIC product won't see the light of day for another 6 months, if at all, which doesn't give a lot of mining time in our shrinking 50BTC/block window.

[fred0]

Where exactly is that ASIC again?

Do you want to know? I'll be glad to tell you: [Tycho] AKA Deepbit is developing it. You're welcome.

Don't tell me: another 1.05GHash/s, 20W, $499 preorder miracle box? Seems we're coming full circle.

Realistically, further FPGA/ASIC development is inevitable as long as bitcoin mining remains profitable. We can look forward to another 12 months of 50BTC/block reward, but after that the reward is cut in half: 25BTC/block. Few expect the BTC->USD exchange rate to be affected by this much, which will leave a lot of the existing GPU crowd in unprofitable territory if BTC prices are in the $2-$4 range by then. FPGA and ASIC miners are the only real option, but whether they are viable depends on their ROI.

Few will be willing to jump on the FPGA/ASIC train if the ROI isn't attractive. The resale opportunities for an FPGA/ASIC miner are extremely limited since they are very much a niche product; whereas GPUs can be resold at attractive prices without issues.

BFL's 'product' (a loose term, granted, until they actually ship) has a current break-even time of 10 months. Similar to a 5870 or other efficient GPU. What could you resell the BFL box for after 10 months, with no warranty? Half price? I can buy 2 5870s for $400 today and be confident that I can resell them for 10% less ($360) in 10 months.

Even if a newer FPGA or ASIC design achieves 1/10th the power draw of the BFL product at similar hashrates and cost, it doesn't have much advantage from an ROI perspective: the payback for such a device (even assuming 0 power draw) would be ~9 months compared to BFL's ~10 months. BFL's product will be available (theoretically) in month. Any other vaporware FPGA/ASIC product won't see the light of day for another 6 months, if at all, which doesn't give a lot of mining time in our shrinking 50BTC/block window.

Sorry Epoch,

You are using logic. Not spoken here.

[defxor]

Few will be willing to jump on the FPGA/ASIC train if the ROI isn't attractive.

There are a lot of people willing to pay out of pocket to effect changes they see as positive for society. These people aren't interested in running hot noisy boxes in their apartments, but a cool low power draw and silent little box is something they're willing to fire up and forget - if it has a chance of upsetting a (bad) status quo.

No ROI, at least monetarily, involved.

[rjk]

Don't tell me: another 1.05GHash/s, 20W, $499 preorder miracle box? Seems we're coming full circle.

No stated ghash, no stated wattage, no stated price, no pre-order (that I know of, yet).
However, also no website or announcement either.
Just wait a few months, keep your money to yourself, and watch the magic unfold.
Lots of things are possible when you have plenty of money.

[bulanula]

Don't tell me: another 1.05GHash/s, 20W, $499 preorder miracle box? Seems we're coming full circle.

No stated ghash, no stated wattage, no stated price, no pre-order (that I know of, yet).
However, also no website or announcement either.
Just wait a few months, keep your money to yourself, and watch the magic unfold.
Lots of things are possible when you have plenty of money.

So it is indeed them that have been unloading the BTC on Gox to fund this crap ASIC ? No wonder.