Quote from: Golempl on January 17, 2018, 10:40:18 AM

If you need any more details let me know.

Thank you for the log. We have a clue why this might be happening but we will send you a new version via PM later today to confirm if the workaround is working.

Quote from: Marvell2 on January 17, 2018, 02:04:15 PM

Some rigs crash as well. Do you want copies of the log files ? This happens on claymore too but faster on this miner , and not even after a disconect and reconnect

Unfortunately the logs won't provide much information as we already report any OpenCL errors in red color directly in the console window. If you don't see any OpenCL errors in the console window, then all OpenCL calls seem to complete successfully. This is definitely a driver instability because of the high memory or core overclock (or too aggressive memory straps even at normal memory clocks). The only thing that can be done to alleviate such problems (besides scaling down the overclock just a little) is to add couple of -minspeed like options: one for the whole rig and the other for individual GPUs as described bellow. If the hashrate of a GPU falls bellow the specified threshold for more than minute or so, the miner will restart itself (or the whole rig if you specify -rmode 2).

Quote from: darkmusika on January 18, 2018, 12:02:16 AM

Whoah! Would really love to have this option.
Can't wait for it

The plan is to have -minRigSpeed option that would be equivalent to Claymore's -minspeed, and -minGpuSpeed. On the second one if you specify one value, it will apply to every card and if you specify multiple values they will be used on card-by-card basis and apply different min speed for each GPU (if you don't want one some of the GPUs to be monitored for min speed you can specify 0 for their min speed).

Quote from: twotwosix on January 16, 2018, 09:16:03 PM

Awesome work, i Hope you can add commands -cvddc -mvdcc -cclock -mclock for control the gpu.

They are included in the TODO list but we need to fix some other issues first. We hope to start implementing the AMD clock/frequency controls in two-three weeks.