Quote from: Claymore on May 09, 2018, 08:28:18 AM

Quote from: Ursul0 on May 09, 2018, 08:18:00 AM

Quote from: Claymore on May 06, 2018, 04:35:42 PM

I'm fine, thanks, is v11.7 too old and you want v11.8 right now?

Maybe...
since the new kernels for polaris (around v 11) I've started to experience application crashes (maybe once in a week across some dozen workers), which never happened to me before. my guess is that this new behavior has to do with the driver.(I'm mostly using 18.2.2) likely occurs with higher mem overclock
I have hw watchdogs so the machine would restart if the system hangs, so the new versions actually exhibit degraded performance, comparing to the previous ones in this regard.
I can configure the watchdogs to monitor for a certain process(will have issues switching algoes), or finally setup Chocolatey:)
But figuring out what exactly is wrong would be of course much nicer.

You can use "-oldkernels 1" if you don't like new kernels. A lot of modded BIOSes and OC settings were made for old kernels so new kernels can cause problems if you don't readjust these settings, "-oldkernels 1" option is the solution for this case.

the performance improvement on the new kernels is very nice...so...
I did adjustments where it was required switching to the new kernels... currently on 70% of my rigs(mainly 12 GPUs setups) runtime is 200+ hours (latest reboot 16 hours ago), still once a week one/few may crash, and I'm not sure I can do more "fine tuning" as it's not clear what the problem is...
Also it seems there's higher probability of miner application crash if xmrig cpu miner runs alongside Claymore.