Status as of Monday, September 8

Discussion of development releases of Prohashing / Requests for features
Forum rules
The Development forum is for discussion of development releases of Prohashing and for feedback on the site, requests for features, etc.

While we can't promise we will be able to implement every feature request, we will give them each due consideration and do our best with the resources and staffing we have available.

For the full list of PROHASHING forums rules, please visit https://prohashing.com/help/prohashing- ... rms-forums.
Post Reply
User avatar
Steve Sokolowski
Posts: 4585
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Status as of Monday, September 8

Post by Steve Sokolowski » Mon Sep 08, 2014 9:10 am

This was a very productive weekend, despite my losing a few hours to a Toastmasters event. Here are the things that have been released and are already running:
  • Lost blocks cause invalid timestamp too early fixed: some blocks were being lost because the timestamp of the blocks was too early. Upon further investigation, I determined that this issue occurred when the server was mining one coin, a switch to a second coin occurred, and then we switched back to the first coin before a new block occurred on the first coin. We would still use the earlier timestamp instead of the current timestamp, which the networks thought was too old (it apparently has to be greater than the average of the last 11 blocks).
  • Lost blocks cause duplicate work fixed: when the pool finds a block, it takes about 250ms for the coin daemons to determine whether the block is valid or not. Previously, since I didn't know that, miners would continue mining the previous block until we got notification that the block was accepted (or rejected). The work performed during that interval would be wasted, because we were mining at the same height, so only one of the blocks would be accepted. Now, upon finding a block, the entire pool is switched to the second-most profitable coin until we hear back from the most profitable coin. While that causes ASICs to lose some hashrate because they are inefficient at starting and stopping work, not switching would cause wasted work, and therefore wasted electricity. This fix improves profitability up to 10% for fast coins.
  • Easy coins being mined with difficulty too high fixed: previously, we had to set a minimum difficulty because the pool could not send easier shares to miners when the network we were mining was too easy. Now, if the network difficulty is less than the pool difficulty, miners with powerful ASICs will be told to pound away at the easy coin at a lower difficulty.
  • Added hot configurable difficulty settings: the pool can now be configured to have a different maximum and minimum share difficulty while running. Using Kristof's ASICs, I tested the pool's performance at a difficulty of 2^14, and determined it worked fine, so that will reduce bandwidth usage for miners by a factor of 4. I might raise this difficulty to 2^15 or even 2^16 at some point.
Here are additional things that are fixed but which have not been released:
  • The pool now displays the amount that miners have ever earned for each coin.
  • There are now sliders below the "daily earnings" charts, on both the "charts" page and the "earnings" page. Users can scroll to the left to view their electricity usage and mining earnings at any point in the past, whereas the current release has earnings fall off the back at the end of a week.
Those website features have not been released because Chris has not performed the level of testing we require to put things in production yet.

The next thing we need to work on with the mining server is researching one final cause of lost blocks "hashMerkleRoot mismatch," which is keeping some coins from being mined. We then plan to add more coins in preparation to add Poloniex at the end of the month, which will provide us with a greater choice of coins, and therefore a greater chance of mining the most profitable coin. At this point, I would think that we are ready for testers to bring the pool up to 500Mh/s or so, as I'm confident that the pool is stable, even if it is not as profitable as it can be yet. Chris's installation of the new router reduced the number of orphaned blocks in addition to these changes in lost blocks, since we now have up to 60 connections on most coins compared to the 10 we had before.

The changes this weekend brought the pool's profitability to about 0.001BTC/day, the highest of any pool. The current 52 cents is the highest we have had since the pool's inception. Feel free to offer comments if you have them.
Post Reply