LTC scrypt difficulty factor 65536 ( for nicehash )

[ghost]

Hi Forrest,
thank you for your great software! i have a problem:
I am renting big hashrates at cloud hashrate providers like nicehash.com

p2pool + btc is no problem.. the username

1GABCZjGVVMBca3puzUCs7occhV2YsY/999999999999+999999999999 

sets a difficulty of "999984.7412" for my worker .. so statistically we will need about 114 shares of the worker to get one p2pool-share of 11463938.22 .. so very big hashrates are possbible..

but when i use run_p2pool.py --net litecoin , the username

LiZBSYYaXcnBF1tdWzSYAqD2H1iAzAGzMq/9999999999+9999999999

leads to a difficulty of 0.2328 .. so we will need 12 shares to reach the p2pool-share-diff of 2.73 .. but at scrypt, all difficulties have to be multiplicated by 65336 .. normal pools will tell the worker a difficulty of 14749 ... i know that this does not matter statistically.. the client now simply creates 65536 *12 = 786432 shares to reach the p2pool-shre-diff of 2.73 *65536 = 178913. but this creates very much packages / traffic.. and probably 65535 of every 65536 packages have a hash > target error, just because the miner and p2pool have different opinions what the diff should be.. i can see that in my logs...

2015-01-08 21:33:52.462176 Worker LUfGskrLkhmmqgsT2AGiN23uD7cjc4uWJk submitted share with hash > target:
2015-01-08 21:33:52.462376     Hash:   352546cd0aaf22c968fab60f051e533805ece20cbf60da0e19b6562638
2015-01-08 21:33:52.462532     Target: 44b82fa09b5a52cb98b405447c4a98187eebb22f008d5d64f9c394ae8
2015-01-08 21:33:52.473330 Worker LUfGskrLkhmmqgsT2AGiN23uD7cjc4uWJk submitted share with hash > target:
2015-01-08 21:33:52.473443     Hash:   e134ce877affe0614c67f044eeba43ca1d8feb92a13c2e8a7f2c444fd4
2015-01-08 21:33:52.473548     Target: 44b82fa09b5a52cb98b405447c4a98187eebb22f008d5d64f9c394ae8
2015-01-08 21:33:52.520353 Worker LUfGskrLkhmmqgsT2AGiN23uD7cjc4uWJk submitted share with hash > target:
2015-01-08 21:33:52.520433     Hash:   b609cac916d6e16cad2f2d00b4922d85f64f0d695899647a3e16a5a251
2015-01-08 21:33:52.520485     Target: 44b82fa09b5a52cb98b405447c4a98187eebb22f008d5d64f9c394ae8

i once wrote a proxy which modified the stratum messages.. when i multiply the number here:

{"id":null,"method":"mining.set_difficulty","params":[0.2328]}

to

{"id":null,"method":"mining.set_difficulty","params":[14749]}

everything will be fine and p2pool and the miner are in sync

i also found in the config file p2pool/p2pool/bitcoin/networks/litecoin.py

DUMB_SCRYPT_DIFF = 2**16

but this obviously will not work..

am i missing something? p2pool is mining ltc for a long time now.. i wonder that nobody had this issue before..

here my config;

python run_p2pool.py --version
13.4-57-gdfad3cd
python -V
Python 2.7.6
uname -a
Linux f634 3.13.0-43-generic #72-Ubuntu SMP Mon Dec 8 19:35:06 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

[ghost]

oh i think you changed it once because of this.. #202
so here is the answer to myself:
we can fix it with changing p2pool/bitcoin/stratum.py line 44 from actual

self.other.svc_mining.rpc_set_difficulty(bitcoin_data.target_to_difficulty(x['share_target'])).addErrback(lambda err: None

to

self.other.svc_mining.rpc_set_difficulty(bitcoin_data.target_to_difficulty(x['share_target'])*self.wb.net.DUMB_SCRYPT_DIFF).addErrback(lambda err: None)

yep.. the hash > target errors are gone now!

[CohibAA]

@buechling - thank you for posting this! This fixes the messages.

[dtlgc]

Would be nice if this was pushed into the main git - just fixed my problem in 2017?

thanks

[farsider350]

I have made a Pull Request for this and am closing the issue. Thanks for the contributions.
#413