Lolliedieb / lolMiner-releases

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

1.86 - GRAM/TON dual mining infinite recalculation loop

aiwash opened this issue · comments

Hi, Win11, RTX4070

image

I always have that message "Restarting dual tune to significant clock change" while running the miner.

.bat file:

lolMiner.exe --algo FISHHASH --pool de.ironfish.herominers.com:1145 --user XXX --dualmode TONDUAL --dualpool ton.hashrate.to:4002 --dualuser XXX --devices 0 --ton-mode 6 --cclk 2010 --moff 2000 --pl 210
pause

The moment the turning starts, I have the defined clocks. After the end of the tuning, GPU loses around 20% clk speed and that restarting dual tuning in the infinite loop.

Single coin mining is working as intended.

So I left it mining for some time, the GPU is very cold, looks like it's not mining at all.

Please help.

Use new 1.87. Problem is that you are limited by the Power Limit. Don't use Power Limit. Use --cclk values

I will try to update the miner and get back to you. Removing GPU configuration from the miner is not solving the issue.
btw, the arguments are taken from the "hashrate.no"

The loop is because something block thr cclk that is for the PL been Touched

Updated to version 1.87. Removed the power limit.
But the clock optimization got to be less often (every 2 min, not 30s). After the "dual mine optimizations", the clock drops from 2010 to around 1600.
image
Meaning, it did not help.

As I told you... clock changes because your config touch the power limit. If that happens will retune. Reduce --cclk value or remove pl

I removed the power limit and clock speed reducing is not helping either, it is happening in any clock speed. Before setting it in lol miner I used Afterburner.
Once again, no issues with single coin mining.
Or you asking me to set 1400cclk? for 4070 with the 2550cclk out of the box it makes no sence.

Try --cclk 1400 and only mining not using others apps. That will not change the clock. And will not retune. You can also block a factor always with --dualfactor value

Well, the clock fixed on 1410, but it looks useless with 2kMh. I must play with the dual factor value indeed.