Quote from: efudd on November 07, 2018, 06:05:30 AM

Quote from: Marchcat2008 on November 07, 2018, 05:52:12 AM

Quote from: efudd on November 06, 2018, 09:12:25 PM

Quote from: Marchcat2008 on November 06, 2018, 08:12:02 PM

Jason.
Can you see my request? https://bitcointalk.org/index.php?topic=5036968.msg47606596#msg47606596
Thank's.

I've started collecting data on what is going on with nicehash. I'm going to have to contact them because I do not see obvious explanations for the failures. extranonce is supported by this firmware, it just does not function properly sometimes. I have just submitted the traces to nicehash to try to get some assistance in understanding why this works sometimes, and fails sometimes.

Thank you,

Jason

Maybe this information will be useful: https://www.nicehash.com/help/patches-for-extranoncesubscription?lang=en

Yeah, I've been through that, and the patches. The extranonce stuff actually appears to be working just fine unless there is something wrong in the calculations of the returned results. I have two traces that are exactly the same (as best I can tell) except for the work requested to be performed. nonce's the same, etc... one works, the other doesn't on NH.

When I connect to some of their stratum, they provide extranonce information back, on some they don't.... I may do a set of tests tomorrow without #xnsub and see what happens.

Here is an example -- one of NH's servers did provide me an extranonce setting:

{"id":null,"method":"mining.set_extranonce","params":["0000908e00"]}

Yet that format does not match their specification. It should have a second parameter option of '5' per their own documentation... I just don't know at this point in time.

Thank you for providing the link!

Jason

Now I am connected to the nicehash pool without #xnsub. With such a connection, rejected are much smaller than with #xnsub. And even with so many deflected ball, the profit is greater than when connected to the flypool. By the way, when you connect to the Japanese server, reject much smaller..