bitcoin p2pool
Total 6.80GHash/s --- Dead 1.15GHash/s
Total 18.8GHash/s --- Dead 7.66GHash/s
litecoin p2pool
Total 18.7MHash/s --- Dead 3.03MHash/s
Total 29.8MHash/s --- Dead 10.7MHash/s
how to reduce dead hashrate?
starting from what I think is most likely
1) latency from the majority of p2pool's hashing power to your node
2) latency from your p2pool server to you (<--- and now that i think about it, that would be king. you wouldnt get dead shares from other nodes being slow to you, those would be orphans)
3) running p2pool on overloaded/underpowered server and/or bandwidth issues (most likely upstream)
4) general bitcoind issues
5) stratum
some ways to fix
ed (adding something and moving things around for more dead share oriented instead of orphans)
0) better internet connection
1) reduce maximum block size, try to do at least 3-5k or so. there was some other block with a +25 fee earlier today that someone with maxblocksize 1000 would have missed. with dead shares and not orphans, this would probably only help if bandwidth was an issue
2) optimize tcp, maybe some tool like tcp-optimizer in windows, ethtool/ifconfig in linux (netsh int tcp show global etc, congestion provider ctcp, all this stuff)
3) if using cgminer, use "fix-protocol" : true, in conf file, to turn off stratum
4) put bitcoind on a ram drive if you have enough RAM (or if you can't handle the whole 9GB or whatever, set dbcache to 2GB or so maybe), also can put some nodes on there with addnode
(other stuff that probably wont help with dead shares)
5) if bandwidth isn't the problem, increase the # of default outgoing connections in the source
6) add some "good" nodes with --p2pool.node xx.xx.xx.xx when running p2pool
if nothing helps, try someone else's p2pool to see if it's any better. if not, then you're probably better off using some other pool, rather than p2pool w/ its 6 resets a min
