Two mining rigs are threatening to disrupt the Ether mining community: the recently released Bitmain Antminer E3 and Innosilicon's forthcoming A10 Ethmaster. Some believe these machines might mark the end of GPU mining. Others don't seem so certain.
One user posting on the Fellowship of Ethereum Magicians website, BsDum, estimated that "effectively, an A10 ASIC is equivalent to the performance of 2.6 best-tuned GPU rigs … and at 50% of the operating costs." The precise numbers are hard to pin down, as GPU rigs are well-known for being customizability, but regardless of a community member's stance on the issue, most seem to agree that ASIC mining is much more powerful and efficient than GPU mining. However, the investment cost is significant: The A10 runs for about $5,700.
This isn't a new concern. Back in March, Piper Merriam of the Ethereum Foundation and Trinity posted Ethereum Improvement Proposal (EIP) 958 to GitHub to modify block mining for ASIC resistance. In other words, he proposed a hard fork to make ASIC mining rigs incompatible with Ethereum. He asked for a simple thumbs up/thumbs down vote on the issue: "Should we hard fork to make ASIC mining harder and to demonstrate a willingness to hard fork any future ASIC based ethereum mining[?]"
As of reporting, the EIP fork has 1079 upvotes and only 45 downvotes. It appears the idea had a lot of traction.
A day before Merriam's GitHub post, Vlad Zamfir posted a Twitter poll asking the same question. Out of almost 7,000 votes, 57 percent of participants voted "yes."
And yet, the discussion seemed to have fizzled out until recently. On Monday, on the Fellowship of Ethereum Magicians website, CryptoBlockchainTech opened a thread to reintroduce the hard fork proposal ahead of the coming Constantinople hard fork set to arrive in mid-October. Their argument is that with Ether prices as low as they are, and computational difficulty having been pushed up by powerful ASIC rigs, GPU miners are already barely breaking even. The author writes:
"It seems this would be the best time to implement a POW algorithm change along side [sic] the reward reduction and difficulty bomb removal. I know you would hands down win over the mining community if you finally took serious the amount of profits ASICs are costing the community and forked them off."
They go on:
"I participate in a lot of wall street conference calls and your hesitation and lack of urgency to roll out PoW/PoS with Casper FFG ahead of Devcon 4 curtailed a lot investment [sic] enthusiasm. It is at this moment in time that your team should double down and pour every last scrap of energy into ensuring Ethereum can complete the goals they have committed to this year. Since the 7/13 call when you talked about delaying until 2019, the ETH/BTC pair has lost 40%."
CryptoBlockchainTech appears to represent the faction of the Ethereum community that is most concerned with profits and the value of Ether, rather than the development side of things, but their stance is also held by many in the developer and user community, as is evidenced by Zamfir and Merriam's interest in the subject, and the apparent overwhelming support of an anti-ASIC hard fork.
However, just whose interest such a fork would be in is not entirely clear. There are pros and cons to ASIC mining rigs. ASIC rigs require less energy and have extremely high hash rates. They're smaller and more profitable in the long run. On the other hand, they are expensive to build and purchase, non-upgradeable, and are only made by a small number of manufacturers.
BsDum explains the Ethereum community's concern:
"My greatest fear is that if we don't have an answer to the ASIC problem, come issuance reduction time, we'll lose a significant number of GPU miners, be left with a less secured (less overall hash rate) chain, and very centralized hash rate sources. Coupled with deteriorating overall price sentiments, we might be opening doors to potentially hostile counterparty actions, such as 51% double spend attack [sic] on prominent exchanges, or worst, a direct 51% attack to topple everything we love and built atop of us."
But not everyone is so convinced that ASIC is going to kill mining profitability and centralize the network.
When rumors about the existence of Innosilicon A10 emerged in July, sparking discussion on reddit, many users seemed skeptical that the new ASIC rig would present much threat. Many commenters pointed out that GPU machines aren't actually that far behind ASICs in hash power, and that there are many advantages to GPU mining, such as upgradeability. Others argued that the coming move to proof of stake will make the debate a moot point. To that, at least a few respondents pointed out that proof of stake has been "coming" for years, and has yet to arrive: ASIC rigs are here now and need to be dealt with.
On the Ethereum Magicians thread, one commenter, AtLeastSignificant, wrote:
"There will always be miners. If it becomes so unprofitable that 50% of miners leave, the difficulty will drop and the remaining 50% will be in profit again. That's just the way this works - there's never a risk of the network going offline due to lack of miners because it will always be profitable at some point."
He went on to state his support for anti-ASIC measures, though, "because anti-ASIC efforts were part of the original spec of the [Ethereum] protocol."
In any case, it may be too late to make such a significant addition to the Constantinople hard fork. Due in October, it's only about two months out at this point, and any proposed changes need testing. But that isn't stopping from people pushing for the topic to be included in the next core devs call on August 24.