search

MultiPoolMiner / MultiPoolMiner

Monitors crypto mining pools in real-time in order to find the most profitable for your machine. Controls any miner that is available via command line.
GNU General Public License v3.0
898 stars 343 forks source link

Claymore benchmarks not being committed to stats. not enough time to finish/record benchmark? #390

Closed alucard20724 closed 7 years ago

alucard20724 commented 7 years ago

i have between a 50-75% failure rate of benchmarks not being recorded for claymore in the current masterbranch. I run about 20 benchmarks from 10-200 for -dcri for each algorithm and most of them are not properly recorded and show up as 0 when recorded improperly

alucard20724 commented 7 years ago

further analysis: multiple problems with this one. first one, I had to remove Nicehash from my pools. I was getting massive rejects on decred from nicehash. seems to work fine with other pools. this probably happened with the other dual algorithms to from nicehash, but i didn't investigate further. Removing nicehash from the pools during benchmarking solved that problem.

Second if i set the interval to 180, "-interval 180", then the benchmark appears to complete adequately. (i noticed this with some of the other algorithms also. ewbf had the same issue, and 180 solved it.

alucard20724 commented 7 years ago

just noticed that some of the benchmarks that were properly recorded (decred and lbry) have been dropping off. I looked into the stats and noticed that the 1,5,10,hour were zero, but the day and week still had the proper benchmark.

fonyo commented 7 years ago

I have also noticed the same symptoms with claymore and I thought it was due to my rigs being unstable while pushing the dual-mined coins too far. I went as far as 500 dcri with smaller increments for Decred and LBRY and for both ethash and ethash2gb versions. Benchmarking took hours and more than 50% of them showed incorrect hashrate. I resolved the issue by deleting appropriate stat files for each and one by one to rebenchmark correctly.

I also reported nicehash giving only rejects for decrednicehash as well. #380

I believe miner configs disappearing from the list is a known issue and will be fixed in the next release. In my case, EWBF was doing this with one or two solvers randomly (decreasing hashrate recorded before disappearing). I did have issues of this kind on all my rigs with my fork of v2.6.0.2 but I only had to switch back to v2.5.2 on my 1070 rig which was just killing all the miners slowly ending up with an almost empty list after just less than a day of use. Not sure how this last report can contribute and I cannot reproduce it due to benchmarking taking too long. #387