About using 1060 3GB cards to mine ETH: Well, they're actually still obtainable, since everyone else is buying RX 580s and GTX 1080s. But also salient is that the power consumption is much lower, as posted previously. (Actually, without overclocking, 19.8 MH/s is achievable at 70W.)
Even mining with 1060s has taken off: articles from 6 mo ago show prices for the 6GB cards as being sub-$200, and now they're closer to $300. Even 3GB cards, which may be getting close to outmoded for the increased DAG file size at more recent epochs, are selling for ~$230.
OC'ing Nvidia cards on Linux is a hot mess. Everyone appears to be using Windows, but I'm too cheap to pay MS, and I don't really want to pirate it. Going from ~20 to ~23 hashrate would be nice, though. Some hacking of the xorg.conf file is necessary to enable memory OC; the problem is under Ubuntu, gpu-manager runs and rewrites xorg.conf upon reboot. When I effectively disabled that using "sudo chattr +i /etc/X11/xorg.conf", the system wouldn't boot into X. Then I had to figure out how to get into the Grub startup page (old docs say hold down shift, but I found that Esc is necessary these days) and get into single-user mode to restore the attribute. At least power target can be adjusted down without coolbits.
On my laptop I had another guide open that I realized I hadn't fully tried. Need to give it a shot.
It was amusing to read about how mining pools calculate "shares": by having the miner submit work and verifying that it meets a standard of difficulty below the network standard for the sole purpose of computing the miner contributions. Of course, a given share may be the block solution, but the miner doesn't have enough information to know. Furthermore, at least if we were talking Bitcoin, the coinbase / generation transaction's address would be the mining pool's. Of course, a malicious miner could return bogus data and cause blocks to be missed....
Don't forget the PPLNS pool algorithm used by nanopool (explanation). Pay Per Last N Shares introduces randomness in the algorithm; if I understand correctly, the pool pays based on some recent subset of shares submitted by each miner when a block is found, as opposed to all the shares, as in PPS.
CoinWarz's mining profit calculator is more detailed than CryptoCompare's calculator. The former is more optimistic than the latter, though: 2.928 vs. 2.78 based on the same hashrate, and is still optimistic on the USD profit, even though the former also subtracts out the pool fees.
Regarding hashrate of 1050s, it appears they're 12 MH/s stock, as opposed to 19-20 with 1060s. The 2GB model is ~$125, the 4GB ~$190. Of course the 4GB is in greater demand. 2GB is definitely too small for ETH; perhaps it'd be usable for other alts like ZEC? With one of those miner boards that allows for hooking up >10 cards, that'd potentially be more cost-effective, if not power-effective.
As far as the Ethereum switch to PoS, the FAQ on GitHub describes what that would look like. Apparently there are two types under consideration: chain-based and BFT-style. BFT is "Byzantine fault tolerance" that handles consensus-seeking in the presence of malicious actors. Amusingly enough, the Wikipedia page notes under examples of such failures that this pertains to swarming bees:
A similar problem faces honeybee swarms. They have to find a new home, and the many scouts and wider participants have to reach consensus about which of perhaps several candidate homes to fly to. And then they all have to fly there, with their queen. The bees' approach works reliably, but when researchers offer two hives, equally attractive by all the criteria bees apply, catastrophe ensues, the swarm breaks up, and all the bees die.My hobbies are commingling! (?!)
As far as using Radeon RX 580s, most of those seem to be sold out at under a $400 price point. Amusingly, Newegg has an Asus 4GB card for $280 in stock: "limit 1 per customer" (lol). I'm less enthused about using those because OC'ing them involves tweaking their BIOS.
I'm not sure why I get like this every once in a while, going full-bore into something different. Hopefully in this case it'll be profitable rather than just another money sink.
No comments:
Post a Comment