Hi,
realbminer!
your miner is very cool. It supports 4gb gpus and very stable. restarts only for update to new version
it gives me ~95% accepted shares on sparkpool on all rigs (13x 106/104/1070Ti - HiveOS). while *another* miner run with almost 100% accepted with a bit smaller HR.
but on pool your miner achieves better HR afterall. LA also much lower - about 2.2, while *another* 6-8 up to 20 making some rigs unstable.
I tested grin@nanopool for some time and found that accepted share rate there is better > 99%. (they have lower diff than sparkpool)
on sparkpool most shares are rejected with {"code":10,"message":"invalid share: share context is invalid!"}
(just test on Luxor: Efficiency - 100.00% on all rigs!)
I played with intensity option - and did not find any real improvements. it can slightly reduce/increase cpu usage by -0.5 / +1 and report higher HR, but on pool it was always lower in long time run.
what this option really do?
recently tested 1080 on windows (Oh, mining on windows is pain!) and noticed interesting thing - your miner produces 85% "bus utilization" while *another* just few percents. what is this?
Improved the miner fidelity by 8%-10%. You will not see big local difference but you will see roughly 10% higher speed on the pool side.
can you explain what fidelity is and how does it affect HR on pool side?
afaik, pool accepts shares with appropriate diff. more shares - more profit. or I am wrong?
some suggestions and wishes:
- report accepted shares time
- accepted/rejected shares per gpu stat
- fix reporting 1C / 1% / 1W sometime for some gpus. (use 0 instead, please. 1 is confusing).
- maybe option for set reporting interval? on big rigs with high LA getting gpu stat can produce even higher LA (irq/kworker and etc...)
- support amd for grin someday
Best regards