The privacy-oriented Grin token held a developers’ meeting where they agreed to hold off on making changes to planned proof-of-work updates for the foreseeable future.
In a move intended to give ASIC manufacturers a return on investment on the chips they have already built, a previously scheduled phase out of Cuckatoo32s chips will be delayed until past 2021.
After a public turnaround on the viability and inevitability of single chip ASICs – citing the improvements to heat density, lowered upfront costs, and potentially reduced electrical costs – of the technological usurpation, Grin is no longer supporting the mass-market miners that were compelled by market forces to run GPUs.
The team had originally pledged to keep single chip ASICs “at bay,” but now that Grin-specific ASICs, differentiated by the SHA256 ASICs, have advanced to a position of market dominance and affordability, the team has committed only to eliminate Cuckatoo31s from the roster.
“In short, preventing single chip ASICs no longer seems worthwhile or feasible, but an earlier version of me thought it was, which had led me to the phase-outs,” wrote Grin developer John Tromp.
Three foundries, Samsung, TSMC, and Intel should be able to produce the increasingly efficient Cuckatoo32 ASICs, said Tromp, whose backing hopes will provide the manufacturers with the confidence to continue production.
Tromp is “90 percent” certain the C32s are ready for deployment, but noted if “in [the] worst case, primary graphrate will fall, and GPU miners will be happy to pick up slack with secondary.” All that is certain is that, “c31 phaseout is a foregone conclusion.”
He wrote to , Cuckatoo31s, “currently have a weight of 2^8*31=7936, will see this weight linearly reduced starting around Jan 15, 2020, over the coarse of 31 weeks, to 2^8*0 = 0, at which point it’s effectively gone, leaving only C32.”
The developers also noted that in time if reason arises to decouple Cuckatoo31s, a simple code upgrade will enable Cuckatoo33s to supplant the outdated technology. Though this decision would also be delayed by an 18 month period of deliberation.
Prominent developer Yeastplume said, “First, as you know, 2.0.0 is just around the corner, which is our first scheduled hard fork (or you can call it a ‘network upgrade’ if you prefer). Fortunately for our current situation, this is a forced upgrade, which means that all users of Grin will have to upgrade their software to the 2.0.0 release.”
Yeastplume’s efforts to ensure adoption come after months of incompatibility within the network caused by a previous update that did not receive total adoption. “For whatever reasons many users and particularly exchanges haven’t been keeping up with the latest versions of Grin,” he wrote, which caused miscommunication between Grin and their wallets.
To ensure complete compliance, Yeastplume said that, “all current versions of Grin will stop working as of the HF block in a few weeks,” hoping it won’t come as a surprise when user’s nodes quit.
He also motioned towards coming changes in governance structure, including a request-for-comments period that seeks more community involvement, as well as slower roll outs of planned upgrades.
Additionally, he said the 2.0.0 code is “much better at doing version checks and the software should be better at explicitly informing users of potential incompatibilities, which we hope will greatly help when we do have to introduce new features to support upcoming technologies.