First, I don’t think Bitmain and Innosilicon are competitors to the extent you think they are. They are likely taking turns releasing ASICs to make it seem as if there is competition.
Second, how does the “efficiency” advantage improve the network? All it does is funnel the block rewards to fewer and fewer people. It also squeezes out less efficient miners, which makes the network less secure and less decentralized.
Mining is an arms race and ASICs are atomic bombs. Sometimes, it’s better for everyone if extremely powerful weapons are banned.
Any proof for your last comment that Bitmain and Innosilicon are only fake competitors?
About your second point, you said nobody is more efficient than bitmain, i said that there are more efficient non-bitmain products out and this with proof. It’s another story what improves the network, but this wasn’t your point when i corrected you.
About arms race. Of course, what did you expect in multi dollar businesses? I doubt you can show me a single business where it’s not an more or less armed race, at least i can not think about on. Now that atomic bombs gives your comment some devastating touch, but fits perfectly a hardliner
I said that I didn’t think the guy who is asking for funding will be able to make more efficient ASICs than Bitmain. I didn’t say that other companies couldn’t make more efficient products than Bitmain.
I don’t have proof that Bitmain, Halong, Innosilicon are fake competitors, and I never claimed to. I think that anyone with an elementary understanding of game theory would conclude that ASIC mining is very susceptible to collusion. So my opinion is that they are almost certainly colluding. They would be fools not to. (Even publicly traded companies like Samsung, Hynix, and Micron have been sued for collusion – imagine what it’s like for private companies in an unregulated industry!)
Yes, it’s an arms race. And guess what is a part of that arms race? GPU miners forking to get rid of ASICs. Just as it’s Bitmain’s right to make more powerful ASICs, it’s also our right to adjust the proof of work algorithm to make it more difficult for them to take over. GPU miners have the upper hand because it’s much easier to fork the algorithm than create new ASICs.
I should add that the only reason we’re still debating this is that Zooko – for whatever reason – doesn’t want to fork. The second he agrees to work on tweaking the proof of work algorithms, ASICs are finished.
Maybe, maybe not. I’am not sure if a simple tweaking is enough. There are rumours that the X3 still mines on Monero … If it’s true tweaking might not be that easy as it sounds. Until we do not have full details of how the Z9 actually works and what’s inside i would not bet too much, but you be right.
Ok, after Asics are finished, shouldn’t we just close all the open threads about Asics than? Where is the logic to keep them open after such announcement that Asics are finished? If such stance has taken to keep the Asics out, so it be, but why still discusse it than?
Also, an investment company called 360 Blockchain Inc has invested in ePIC and acquired SV CryptoLab who are using Bitmain’s Antminer in their mining rigs.
From the same presentation, there is an IP strategy mentioned. Is this documented anywhere?
Finally, from the proposal, can a non-profit foundation with no experience in the space really achieve its aims, given the cultural mismatch with competitors and its own partner?
ePIC is a Digital Design house at the core, so we are agnostic to what type of semiconductor the design goes into, GPU/FPGA/ASIC. Given that Bitmain has introduced an ASIC to Z-Cash we assume we would need to develop a better ASIC.
ePIC is based in Canada, our engineering team has left AMD to start the company specifically to build chips for blockchains. Production depends on the design, but we are fabless, meaning we would likely work with Global Foundries, TSMC, or Samsung to fabricate. Distribution/Sales is key question for research in the grant, how do we ensure the greatest possible decentralization of the product?
ASICs is a price to performance business, whoever produces the best design and end product will win out. AMD won the Xbox and PlayStation businesses from Nvidia by making the better product. ePIC plans on winning the war to develop the best blockchain accelerators. How the market is made and potentially controlled is a subject of the research, ie how do we contribute to decentralization of hardware.
Fair enough, I am sorry to offend the home miners and those contributing hash power in pure support of the network. My drive was at industrial miners who are running tens of thousands of machines using 10’s of megawatt hours.
Ok, thx for the answers even not all are answered.
However, what bugs me again is that here we talk again about an Asic, last time an FPGA, bevor again Asic, bevor it was again an FPGA, lol. What is it finally? An Asic or a FPGA??? Seriously, i’am confused and in no way offensive as i’am an one of the minority asic miners without getting love anyway…
Just a little bit off topic. But i got intrigued that you worked with AMD, a gpu producer.
What’s the reason you are now an Asic believer and do not for example work on “tuned” gpu’s for the gpu miners, which of course could be, at least i imagine, a very profitable business as well, maybe even an easier one.
What are the advantages of an Asics vs gpu and where is the future going in your opinion?
Great find ;). We are an ASIC Design House, we develop the core IP design of chips. If Bitmain wanted to work with us we would have the discussion. They are the biggest producer of ASICs in the blockchain market, we would be crazy not to. We, by the way, have not had any conversations with Bitmain. 360 Blockchain is an investor in us and a number of other projects. We have not connected with SV CryptoLab, which I assume is a mining operation, and like most operations runs Antminers.
Yes, ePIC has a really exciting IP strategy, but it is not much of a strategy if we publish it prematurely;).
It is a research project to figure out how a strong open source community like Zcash’s can actively collaborate with and potentially do business with the semiconductor industry. The hardware will happen to you, as Bitmain has demonstrated, or it can happen with you as we are proposing.
By the way, this is asked in other posts but I have hit my comment limit. GPUs are ASICs, they are Application Specific Integrated Circuits for running Graphics. Their advantage is that they are programmable, ASICs in the crypto context are not. Crypto ASICs are ‘dumb’ chips, they do one thing, SHA-256 for example as efficiently as possible.
Our mission is to develop an ePIC product to bridge that gap. Working with the protocols and understanding their needs is how we do that, just like AMD working with Sony and Microsoft to make gaming platforms better products.
Would be interested in your thoughts on Siacoin, who embraced ASICs and partnered with Obelisk to build ASIC miners, only to have Bitmain beat them to market and spoil the party.
Just in the past few days, the community are back to talking about forking and lawsuits, since the Obelisk miners are arriving late and under spec, reducing their profitability as the difficulty continues to skyrocket.
Seems like the pick-and-shovel industry is pretty tough. A non-profit foundation with zero experience in the sector could not only find themselves in a pickle, but also expose the project to dangerous centralization and the removal of grassroots supporters from participating in the project.
It cannot be the Foundation’s responsibility for having chosen Equihash, because the Foundation did not exist when the decision to use Equihash was made. I take personal responsibility for having proposed the 192,9 parameters [edit: I meant 200,9]; @str4d was also involved in that decision.
The reason for choosing those particular parameters (contrary to conspiracy theories that have come up in the other thread) was that at the time we were using an implementation of Equihash in zcashd that, with 144,5 parameters, was taking a substantial fraction of the target block time to produce even a single solution as input to the difficulty filter. This would have caused significant problems in applying Equihash at all, since faster miners would have had an advantage disproportionate to their speed. As it happened the zcashd Equihash solver was missing important memory optimisations, but that became clear only as a result of the open-source miner competition just before the Zcash launch (and too late to safely change the parameters again).
Just to add something to your statement. Baikal and Bitmain were the first there, both released in January, and the Innosilicon and Halong released last month in April.
On Blake2B there are 4 Asics + 1 planned: Baikal Giant B, Bitmain A3, Innosilicon S11 SiaMaster, Halong Mining DragonMint B52. The mentioned Obelisk SC1 is coming late it seems, true.
Interesting is that i wasn’t aware that Sia itself partnered with Obelisk and that these are coming late. Bad business decision?
When did Sia partner with Obelisk. If it’s after January after the first other Asics are on the network allready it’s clear that there wasn’t much of a chance that Obelisk can made it 1 week later only, not?
Thanks for the reply. After rereading my post I can see where I caused confusion.
To try to clear things up:
I was in no way trying to allude that the people who came up with the initial parameters/the development team in general, were somehow at fault or that there is some conspiracy. What I was trying to get across was, the issues pointed out by solar designer were the foundations responsibility to follow up on. Not the development teams. I also think that the foundation should have been continually investigating the link between equihash and hardware. these are my opinions though, and I am more than happy to change them. I dont know what work was done by the foundation in this area, but give the current amount of different asic/fpga vendors it would seem that they went wrong somewhere.
I apologise if I got things wrong and I appreciate your additional information. Again I would like to thank you for your representation in the long thread. Your responses have given me a lot more confidence in the zcash co team. (in fact the willingness of the entire devteam to engage with the community is a great confidence builder)
Part of the foundations mission states this:
We will act as stewards of the Zcash protocol, providing leadership and supporting its continued maintenance and improvement.
we will offer responsive leadership when changes are necessary on short notice, such as a hard fork to resist an attack.
We will help coordinate on upgrades to the protocol. This can include organizing soft forks, organizing hard forks, and helping to conduct “ceremonies” such as parameter generation (should more of those be necessary).
Sia and Obelisk are founded and operated,
I assume, by the same team, which is an interesting strategy that I am keen to learn about from them. The challenges they are having reflect exactly why ePIC exists making top of the line hardware is really hard and needs deep expertise. The cliche that hardware is hard is true. Being first to market is super important but performance is the critical factor. Crypto has an interesting opportunity to figure out how to factor distribution more deeply in the equation in order to achieve decentralization and access. Sia is running an important experiment for everyone concerned with the infrastructure of decentralized networks. IMHO we should be cheering them on.
Halong Mining DragonMint B52, Sold Out, Blake2B, 3.83Th/s, 1350W
(seems like a re-brand only)
While the Obelisk is more effiecient than the Bitmain counterpart A3 it seems both are lightyears behind Innosilicon. The Baikal as a multi algo miner is as well out in Blake2B while still good at least on his other algos).
With this compared, i think the Obelisk has not much of a chance to be honest. Not even talking that it gets delivered in June, 6 months behind the other first Asics on Blake2b, i would go as far as saying that 6+months are lightyears in cryptospace…
David Vorick, who I believe is the founder of Sia and of Obelisk, is coming to Zcon0 to give a presentation. Presumably that will include some good insights and lessons-learned!
