There are a lot of naysayers here on the forum who say it is not possible, but keep in mind the same thing was said about Litecoin. I would say a Proof of concept is possible, but it will require a lot of time and then tons of money to produce at high volume. I've worked in Semiconductors for 3 years and I was also one of the very first to look at Litecoin FPGA poc in 2013, which was successful and then went ASIC. Notice, Ethereum is described as being ASIC resistant. This does not imply an ASIC or FPGA cannot be done.
@chrizzo Mate i got 35 years in electronics. Have you got a circuit design and Gerber files for copper layers, solder mask, legend, etc. all ready to go? Or did you hope someone would work for free? Have you a cheap chip suppler that can scale up as the orders come in? But most of all can you promise better efficiency in ASIC operation as well as production (well better than AMD) let alone driver software design? Have you got someone that can assemble, pic 'n' place, distribute? All at a lower cost than AMD in China?
SO YES.. it can be done but how much have you got to spend cos im not working on a promise and i want a retainer up front too, and I only travel 1st class and you will have to pay for the travel of my close protection operative too.
Also, can you name a current Bitcoin ASIC manufacture that is not Chinese that is still in business? I hear spondoolies tech was made bankrupt last week by an Israeli court . Are Bifury still going?
Hey, we could make a proposal to the DAO! It's got sh!t loads of money now
Sure, but can we build it before the timebomb (shift from PoW to PoS) goes off. i.e. just over 6 months left of mining. We need to have it built and in production by the end of this month. Then we have 5 months to sell and distribute (no one will buy in the last month of mining). I'm up for it but the wage has to be right. Non of this, "Oh we get paid at the end!"
A better move would be to make a PCB with a memory chip that will plug into a USB socket to hold the DAG file rather than use the GPU memory store. Thus allowing anysize GPU to be used to mine. Still only got 6 months to make our money back and turn at least 50% profit.
An alternative route for ASIC would be realtime DAG slice generation. This removes the requirement for lots of expensive high bandwidth memory. Apparently Claymore's GPU DAG generation kernel generates 1.4GB of DAG in a few seconds on a modern GPU. You only need 2KB of DAG per hash, so he could already demonstrate a DAGless kernel at about 100KH/s (not sure how many seconds the DAG generation takes, exactly). Make an ASIC / FPGA design for that and off you go. I wouldn't be surprised if the Wolf0's of this world have been toying with this idea for much longer already.
Comments
But most of all can you promise better efficiency in ASIC operation as well as production (well better than AMD) let alone driver software design? Have you got someone that can assemble, pic 'n' place, distribute? All at a lower cost than AMD in China?
SO YES.. it can be done but how much have you got to spend cos im not working on a promise and i want a retainer up front too, and I only travel 1st class and you will have to pay for the travel of my close protection operative too.
Also, can you name a current Bitcoin ASIC manufacture that is not Chinese that is still in business?
I hear spondoolies tech was made bankrupt last week by an Israeli court . Are Bifury still going?
We need to have it built and in production by the end of this month. Then we have 5 months to sell and distribute (no one will buy in the last month of mining). I'm up for it but the wage has to be right. Non of this, "Oh we get paid at the end!"
A better move would be to make a PCB with a memory chip that will plug into a USB socket to hold the DAG file rather than use the GPU memory store. Thus allowing anysize GPU to be used to mine. Still only got 6 months to make our money back and turn at least 50% profit.