Status as of Friday, March 20

[Steve Sokolowski]

I decided to take the day off today. I won't have to try to get up the street to work in the snow, and I will have 10 additional hours to work on the project this weekend. Here's today's status:

• I've completed writing the first phase of the code that is work restart-aware. The software in testing computes which coin is optimal for a miner to mine based upon his observed work restart interval. Customers will not be required to set this value. However, it will not go into effect until customers have been first connected for up to 15 minutes, because it takes time to determine the value.
• Miners who have long work restart times will find themselves being assigned to more difficult coins, like Monacoins and Litecoins. They will be penalized, and the penalty will be distributed to the remainder of the miners so that the suboptimal coin selection doesn't hurt everyone. However, in all cases the penalty lost will be much less than the amount of money that would have been lost by just mining Litecoins.
• We will need to add new columns to tables for this bonus/penalty. Since our tables have hundreds of millions of rows, that means there could be an hour of downtime when this is rolled out, probably on March 28.
• I contracted with bpj1805 for him to expand our testing software to test this new work restart algorithm.
• Chris has been working with coin developers to add their coins for payouts in exchange for advertising. His strategy appears to be succeeding in bringing in more miners. However, he is finding that, for many coins, the developers are actually extremely difficult to find. They don't post any contact information and don't reply to messages and E-Mails. coinaday, the Nyancoin developer, is an example of someone who is an exception to this rule.

[kires]

I decided to take the day off today. I won't have to try to get up the street to work in the snow, and I will have 10 more hours to work on the project this weekend. Here's today's status:

• I've completed writing the first phase of the code that is work restart-aware. The software in testing computes which coin is optimal for a miner to mine based upon his observed work restart interval.
• Miners who have long work restart times will find themselves being assigned to more difficult coins, like Monacoins and Litecoins. They will be penalized, and the penalty will be distributed to the remainder of the miners so that the suboptimal coin selection doesn't hurt everyone. However, in all cases the penalty lost will be much less than the amount of money that would have been lost by just mining Litecoins.

I'm really looking forward to seeing the efficiency improvements this can bring. From what I've seen so far, a single titan (with no bad dies) runs at just about 340 mhs hashing away on LTC or any single scrypt coin with a block time over a minute or two. While pointing here is much more profitable in terms of Mh/Btc, the frequent switching (more accurately, the titan's crappy handling of frequent switching) means they average about 280-300 Mh/s, which drags the bottom line down to pretty close to straight LTC.

[loszhor]

If one simply voluntarily sets a static difficulty and directly merge mines Litecoins or Monacoin, as mentioned, can they avoid the penalty?

[kires]

I think they would avoid the penalty but they would also avoid the higher profits from mining the most profitable coins, so I suspect it wouldn't be worth it.

[Steve Sokolowski]

To answer loszhor's question, the default will always be more profitable than static mining a single coin. After the update, you will only want to static mine a coin if the purpose of your doing so is to advance its blockchain or secure its network.

The reason is that if it turns out that your miners are so inefficient that litecoins are the best coin for you to mine, then you will be assigned litecoins 100% of the time anyway. Your penalty will always be the difference between the optimal earnings and the litecoin earnings, which in this case exactly equals what you would earn mining litecoins.

However, most people have miners that are at least a little better than 100% litecoins. In that case, you'll be assigned a better coin for your specific miner, and the penalty will be less.

[loszhor]

To answer loszhor's question, the default will always be more profitable than static mining a single coin. After the update, you will only want to static mine a coin if the purpose of your doing so is to advance its blockchain or secure its network.

The reason is that if it turns out that your miners are so inefficient that litecoins are the best coin for you to mine, then you will be assigned litecoins 100% of the time anyway. Your penalty will always be the difference between the optimal earnings and the litecoin earnings, which in this case exactly equals what you would earn mining litecoins.

However, most people have miners that are at least a little better than 100% litecoins. In that case, you'll be assigned a better coin for your specific miner, and the penalty will be less.

Yes, the main reason I ask is that I am testing a contract miner from the Beta Mintsy service and unfortunately the contract miners have disappointing start up times and stability issues especially when changing coins and I suspect difficulty as well which can last up to 5 minutes per switch and in some rare cases up to 30 minutes. The best solution I have found is fixing the my pool parameters as much as possible.