BFL ASIC is bogus

[ElectricMucus]

I can see I'm already way in over my head, as I don't understand most of what the article says. But it seems that my view of hardware design as "write HDL, create chip" is grossly over-simplified.

What you are describing are fpga to asic conversions like Hardcopy.  These chips consist of the same logic layout as the fpgas you use to protoype them and they have metal layers instead of the routing fabric making them faster and cheaper to produce.

You can check out Chuck's site yourself there are some examples about what his software does.

Creating a full custom asic involves modelling the transistors, simulating the behavior of the p-n junctions and the currents through the metal layers. Next is constructing logic elements out of the transistors and configuring the layout of the chip and transferring the logic into it. There are ways to start with pre-made logic elements if you buy the software which models them for you, or rent it. There are also processes which are some in-between where you have some areas already layed out for you. There is lots of snake oil in this area... it boils down to pre-configured layers which cost you performance.

[1714617221]

1714617221

[1714617221]

1714617221

[1714617221]

1714617221

[1714617221]

1714617221

[1714617221]

1714617221

[1714617221]

1714617221

[davidspitzer]

Does it really matter whether it's full custom, off the shelf or in between? If it hashes at the promised level they can send me a paper towel tube wrapped in bacon with a USB port if it works.

[ElectricMucus]

Have you even read the thread?

Does it really matter whether it's full custom, off the shelf or in between? If it hashes at the promised level they can send me a paper towel tube wrapped in bacon with a USB port if it works.

Except it just won't work. Again anything but a full custom asic cannot supply the promised performance. Is this some kind of wear down tactic to always come up with the same question/arguments even when they have been thoroughly answered/debunked?

[Puppet]

Have you even read the thread?

Does it really matter whether it's full custom, off the shelf or in between? If it hashes at the promised level they can send me a paper towel tube wrapped in bacon with a USB port if it works.

Except it just won't work. Again anything but a full custom asic cannot supply the promised performance. Is this some kind of wear down tactic to always come up with the same question/arguments even when they have been thoroughly answered/debunked?

Thats nonsense. The difference between a full custom asic and standard cell asic is relatively small, usually less than one process node, and we dont even know what node they used. For some applications and particularly if you dont have very skilled designers with a ton of time, a standard cell solution might even perform better, but will typically be a bit bigger.

Im pretty sure BFL will deliver a standard cell asic, and their "full custom" refers to the asic design, not the transistor design. A poor choice of words perhaps, but nothing more.

[ElectricMucus]

They promise 3.5GH per 5W (usb powered) they won't even reach that with the latest nanometer node, so the debate is irrelevant.

I was refering to full custom asics as "real" asics to which standart cell asics belong since they use the same manufactureing technique except they use tech libraries instead of construcing their own gates and sub-circuits. On the physical layer it is the same thing. FPGA conversion asics however are not, and that's the only "asic" which BFL would have access to.

[Puppet]

They promise 3.5GH per 5W (usb powered) they won't even reach that with the latest nanometer node, so the debate is irrelevant.

and you base that on what exactly?

[abracadabra]

They promise 3.5GH per 5W (usb powered) they won't even reach that with the latest nanometer node, so the debate is irrelevant.

I was refering to full custom asics as "real" asics to which standart cell asics belong since they use the same manufactureing technique except they use tech libraries instead of construcing their own gates and sub-circuits. On the physical layer it is the same thing. FPGA conversion asics however are not, and that's the only "asic" which BFL would have access to.

BFL has mentioned that the USB power is not for the mining hardware.  It's only  for the coffee warmer.  The circuit will require it's own external power supply.

[beekeeper]

They promise 3.5GH per 5W (usb powered) they won't even reach that with the latest nanometer node, so the debate is irrelevant.

I was refering to full custom asics as "real" asics to which standart cell asics belong since they use the same manufactureing technique except they use tech libraries instead of construcing their own gates and sub-circuits. On the physical layer it is the same thing. FPGA conversion asics however are not, and that's the only "asic" which BFL would have access to.

BFL has mentioned that the USB power is not for the mining hardware.  It's only  for the coffee warmer.  The circuit will require it's own external power supply.

I'm pretty sure they said they calculated they can use the heat generated from the asic for the warmer.

[Cergorach]

*shrugs*

Assuming folks mine BTC for profit, the BFL ASIC is a necessary investment if folks want to stay competetive in the BTC mining market. I suspect that most folks that want to buy a BFL ASIC already have or have ordered the BFL FPGA products (Single or Rig) and those will be exchanged for the BFL ASIC when they become available plus half the price of the ASIC. They can still mine during the time they wait on the new product, sure they paid a bloody big chunk of cash up front, but I assume that as this is an investment, they are prepared to lose that money. That's what business is all about, taking risks and making money. Not taking the risk most probably will cost more money (that is an educated guess because we can't see the future and only hindsight is 20/20).

Let's say that BFL doesn't come close to the state hashing numbers, as long as performance is at least twice what it is now, no money is really lost. Maybe a bit more to compensate for the additional difficulty to compensate for the added extra hashing power. I doubt power consumption would be twice as high, because that would probably generate more heat then can be dissipated. Heck, if a $30k Rig could only do 100GHash/s, that is still four times more then a $15k Rig could do. Sure the advantage over the folks not using the new BFL products would be less great, but the financial advantage is still very much there.

If it was a scam, it could be made far more attractive for folks to invest in, more 'believable' numbers could be used, lower prices so more folks would be willing to risk it, etc. This reminds me very much of the time when the rumors of the iPad 3 and it's very high resolution display, there were many naysayers that it couldn't be done, stating very reasonable 'facts'. The problem was/is that technology isn't an analog increase, it's a digital one, with jumps in performance instead of a gradual increase. Even in the processor industry you don't buy a 100MHz proc and then a 101MHz proc. It went from 100MHz to 120MHz. From DVD to Blueray the maximum amount of data a single sided disk could contain increased 15 times (8.5GB to 128GB), sure those are not mainstream disks, but is a $30k Rig?

Honestly, if you don't trust BFL, your pretty screwd as a big miner (best price vs. MHash/s vs. KW/h). And if you don't trust BFL, that's your right and your welcome to your opinion, because after all, it is an opinion, nothing more. Folks that do believe are also welcome to their opinion. Personally I'll wait and see, not only do I have mixed feelings about the future of the BTC value and the difficulty of mining it, I'm also not currently in the position to mine for profit in any meaningful way (due to not enough budget for a BFL Rig and cooling issues, not to mention very high local power costs).

[ElectricMucus]

They promise 3.5GH per 5W (usb powered) they won't even reach that with the latest nanometer node, so the debate is irrelevant.

I was refering to full custom asics as "real" asics to which standart cell asics belong since they use the same manufactureing technique except they use tech libraries instead of construcing their own gates and sub-circuits. On the physical layer it is the same thing. FPGA conversion asics however are not, and that's the only "asic" which BFL would have access to.

BFL has mentioned that the USB power is not for the mining hardware.  It's only  for the coffee warmer.  The circuit will require it's own external power supply.

Where did they mention that?
Source as usb-powered: The BFL homepage at the time this thread was created, as the consensus of the bitcointalk community. If BFL and their supporters now engage in backpeddeling that doesn't affect my case, it only strengthens it.

A walwart power supply fpga conversion product could indeed be capable of delivering that performance, at least in theory but it is still very questionable if the raw hashes figures would hold at that price.
If you people care to read the thread in it's entirety you'll see that has already been discussed.

[MrTeal]

They replied in their thread somewhere that the actual hashing engine would be USB power, with the external power supply being only used for the coffee warmer.

[runeks]

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

[molecular]

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

Who cares as long as it lives up to the hashrate specs.

[Vladimir]

Whether they will get claimed performance or not is basically a question of whether they have ~30 million USD for R&D and 9-12 month time or not.

[MrTeal]

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

Sorry, there's dozens of BFL ASIC threads and I'm not looking through them all to find the quote.

The basic gist was, it was revealed (I believe on the BFL website, but I could be wrong on that) that there would be a barrel jack on the Jalapeno. People obviously got excited about that and the implications it held for efficiency numbers and the fact that BFL had previously stated the Jalapeno was USB powered. BFL then came in and clarified that the actual device was USB powered and the plug was just to provide extra power for the warmer.

[runeks]

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

Who cares as long as it lives up to the hashrate specs.

People who pay for power, I'd imagine.

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

Sorry, there's dozens of BFL ASIC threads and I'm not looking through them all to find the quote.

The basic gist was, it was revealed (I believe on the BFL website, but I could be wrong on that) that there would be a barrel jack on the Jalapeno. People obviously got excited about that and the implications it held for efficiency numbers and the fact that BFL had previously stated the Jalapeno was USB powered. BFL then came in and clarified that the actual device was USB powered and the plug was just to provide extra power for the warmer.

Thanks for the recap.

[ElectricMucus]

^ I don't understand this. Is the hashing ASIC chip not producing the heat? Please do find the quote if you can.

It seems obvious (from reading this thread at least) that BFL can't achieve 3.5 GH/s at > 5 W. My prediction is that they will release something, perhaps a month or two later than October, but that it will not live up to the power specifications.

Who cares as long as it lives up to the hashrate specs.

Right, and what if that is just a little off lets say by about 40%?
What if next-gen fpgas happen to beat them?

Whether they will get claimed performance or not is basically a question of whether they have ~30 million USD for R&D and 9-12 month time or not.

Psst, everybody knows BFL has a magical contract with Chinese elves.

[YokoToriyama]

Right, and what if that is just a little off lets say by about 40%?
What if next-gen fpgas happen to beat them?

custom or hard copy.. your trying hard to break down the difference between say a turbo and a supercharger..
but come on 40% off., when you have nothing to base this on is really going on speculation.

we can sit here all day and say it might be 50% off the target but right now everyone here
is waiting on 40 Ghash's.. i don't care if its full custom .. hard copy or donuts with a USB cables sticking out of them
40 ghash it is until you come with actual evidence we can go at this till the earth freeze's over.

[ElectricMucus]

You'll see.

The difference between Hardcopy and real asics has been explained several times, just look it up.
One last time: Hardcopy is the same device as an FPGA except there is a metal layer instead of the routing, so there is marginally lower power consumption and marginally higher speed. Its a little better than the difference between different speed grades of the same fpga. Several people many of them professionals in the field have said this if you don't believe me.

btw: nicely done for finding this thread after 13 posts  

[crazyates]

just looking at the network hashrate... something goofy is going on,.  how do we go from 15000 to 22,000 within 24 hour period ?  I have never seen such large swings until lately. 

I thought the "network speed" was based off the rate at which blocks were being found, and estimating it from there. If everyone happened to have a 24 hour streak of bad luck, then the graph would dip. If everyone had really good luck tho, then the graph would swing up. Thats why I like the 3day average.