BBT Livestream Recap: PROGPOW in 2020
BLUF — PROGPOW is a proposed algorithm change to Ethereum. It is considered contentious, as there is not a consensus of opinions (when you view the online collective, blogs, medium post, twitter) to change the current Ethash algorithm. Why am I discussing this right now, wasn’t this old news? Well…
· The second approval to move forward was established in Core Devs Meeting #81 on Feb 21st 2020. The implementation is tentatively scheduled for some time in July (roughly 3 weeks after EIP 1962).
https://www.youtube.com/watch?v=zSRzlC_dCx8
o Specific time stamp added in description (Berlin release)
§ PROPOW Discussion starts — 1:33:00
§ Proposal to split the Berlin (EIP1962) & PROPOW
Key takeaways
· 0.9.3 mining testnet has been live and tested
· Ultimately EIP1962 first, set in June, then 3 weeks later PROGPOW
o Proposal to include in overall Berlin release, but just a switch to enable/disable
§ I.E less overhead for exchanges and clients to upgrade (triggered later on)
o Discussion lead to an action by the dev team to reach loudest folks against PROGPOW, Gnosis founder Martin Koppelmann & SpankChain CEO Ameen Soleimani to validate they would not push for a chain split.
o James Hancock — put his position/name on the line that he felt there would not be a chain split and most of the concern was just that a concern, not real threat to contentions (non-activating, nor chain splitting hard fork)
Now since the very first discussions started on PROGPOW, there has been a range of community involvement to establish some semblance of a decision, not only if, or if not PROGPOW, but if Ethereum should change ethhash or not. This had been done on Coinvotes (carbon votes) and Miner/Pool signaling. Now what’s the Pro / Con of this type of activity.
PRO — Democratic use of Ethereum and ETH as a voting mechanism (your keys, your signature)
- Ironically, requires you to actually vote (do something), click a link, cast a vote
- While sites were created to make this a simple copy/paste digital signature with a Yes/No Flag, the voting was based on ‘stake’ size, which again exhibits part of the issue with the power in voting comes with he who holds the large amount of eth
There have been a few attempts at leveraging blockchain based voting to be as democratic as possible.
- Coin vote of over 3.1 million eth participated in voting. 2.9 million voting YES for PROGPOW and 200k voting no.
- Miner Hashrate has consistently voted for PROGPOW — https://www.etherchain.org/charts/progpow
CON — No vote, no voice; no amount of eth, limited exposure
Why does it matter to GPU miners?
Let’s take a very simple, easy to understand rationale of why any individual in this world care about the Proof of Work mining algorithm. At its most fundamental level, the ability to participate. There is no argument or discussion around the availability of CPU and GPUs, they available everywhere on this planet.
The counter argument may go the way of, well, it ultimately comes down to profitability and does not counter situations of centralization of mining. That is only true when the equilibrium of profitability is near the world average cost of operations vs. current price of the coin. In times of speculative rises, the profit opportunity lets loose for more folks to participate, i.e more will naturally have the incentive to join the network. This includes folks that already have sunk cost will turn back on their investments to participate. Again, reinforcing the simple principle, the ability to participate.
There are active twitter polls and list, having people sign up to say they are against PROGPOW, which is totally cool and fine. They may have their specific reasons and I would venture to propose most of the concern is around the fact a few ‘known’ developers have voiced their particular concern, which makes them concerned. Nothing more, nothing less. That is not to diminish their concern, I believe if enough folks voicing concern that creates a transmission that will reinforce a position. Let me say that again slowly, transmission of a concern, can sometimes be construed to a particular position. The biggest risk in accepting an algorithm change, or (any hardfork) for that matter is a challenge to the chain, which ultimately splits the cryptocurrency into two chains, one ethhash, one progpow. Contentious HF’s do NOT equal Chain Splits, unless miners choose to mine both sides, as was done with Ethereum Classic after the DAO HF.
The signal to noise ratio that I see from the mining community is not to split the currency, not even in the slightest. The contention risk really is coming from the dApp development community voicing their concern that the effort is taking away from other development and that this particular algorithm may lead to a competitive advantage from one GPU company type to another.
I will address each concern separately.
1) The mining algorithm and chain security inherited by Proof of Work fundamentally with respect to dApps is not changing. This is the equivalent in the centralized cloud architecture hat of changing out a mechanism in the infrastructure (IaaS) that the (PaaS,SaaS) participants inherit. Equivalent to an SSL Certificate Type; like swapping DV SSL (domain certificates); not even a type change says moving from a (DV SSL to EV SSL)
a. One of the research aspects early on after the initial dev calls was to estimate the LOE which resulted in a low impact to the development/other EIPs
b. Multiple Testnets were launched and tested 0.9.2 & 0.9.3 spec standards and tested how EIP would activate without issues.
2) The mining advantages around one GPU company vs another (AMD vs nVidia)
a. To test this assertion, I spent over 200 hours of livestreams on twitch to test nearly 68 different GPUs across AMD & nVidia. Results uploaded to google spreadsheet, and quick recap videos provided on YouTube. The results showed an advantage to nVidia but this was NOT due to programmatic rational within the algorithm, but due to GPU architecture and as in games, scientific applications, or raw compute computational jobs, nVidia GPUs were at that time just outright better. This has nothing to do with PROGPOW, this happens to do with architecture at the GPU core level.
i. The ironic part about that position was that the current ethhash was for a long time overwhelmingly generous to AMD for years due to amd’s architecture, memory controller and timing related to the I/O bound nature ethhash has. The nvidia 10 series closed the gap and challenged the position on ethhash with the then first 1060 architecture then 1650s.
ii. AMD has sense made up significant ground on nVidia with its Navi architecture becoming one of the best costs, to power to ethash performance over nVidia cards. Computationally, for the money AMD is beating out nVidia in nearly every category besides the ultra-high end, with nVidia reigning supreme still with the RTX Titan and RTX 2080Ti respectfully. But in a supercar kind of way regarding price and performance.
b. Bottom line, this argument is no longer around the GPU companies, that argument did have some grounds with regards to possibly disenfranchising existing AMD farms, however many have been retooling to newer RTX, GTX 1650 Tis and 5600 / 5700 XT series. Other larger acquisitions come in the scope of low-cost transference of existing Mining Edition Cards such as the tried and true RX 470’s and 480’s from Sapphire.
3) Finally, this brings us to the previous discussion around general angst on changing algorithms. Most of the concern is based on general understanding on how actual mining incentives work and don’t go much past limited assumptions and untested game theory.
- Now that I have covered a general ‘miners view’ Let’s look at the basic situation.
1) Nobody likes game rule changes mid-course, this could include even changing the algorithm.
a. If bitcoin has taught us anything is that industry, will build to reinforce if the rules are clear and a level of confidence exist.
2) The current Ethhash ASICs by and large are actually going to fall off the network, Bitmain devices by mid-April and still unconfirmed reports the A10 will too (4GB of GDDR6) ; so is the existential risk really there for GPUs
a. What I would submit is as POW continues on Ethash, the epoch continues to increase and in the next few years, many GPUs will fall of the grid for mining due to memory limitations.
b. PROGPOW brings back a whole range of GPUs that couldn’t mine anymore, including 1060 3GB cards.
3) Lastly and in closing. Countless cryptocurrencies have changed their algorithms. This is not a high-risk event from a technical challenge. PROGPOW or any other algorithm Ethereum would consider would be done on the behalf to allow more to participate, nothing more, nothing less. It’s not more complicated than that. The dynamics of the way mining incentives then self-correct, this is not a problem for a developer to figure out past that.