Feature Request: Add notification if hash rate changes significantly

[hwengmgr]

I have noticed that when a GPU has a glitch, driver or HW, the reported hash rate goes into the GH/s range when its normally much lower. Sometime you need to just restart the program, sometimes a reboot is necessary. It would be nice to get a notification when this happens, as today you simply have to monitor the machine to see when it does hits change.

If the average hash rate can be monitored roughly, and if a SIGNIFICANT change, like an order of magnitude or more, is seen, assume something bad has happened and notify via email.

cannot reliably repeat the GPU having a problem but the reported hash rate in the GH/s range has also been seen with GPU's where the driver has an issue.

downloaded latest from Github

Windows PC, 1060 GPU

[gertjaap]

Created a separate issue for misreporting the hashrate. We should fix that in stead of notifying the users about it. See #65

Notifying about significant drops/increases in hashrate is a nice feature request. So we can keep this open if anyone feels like implementing that (especially sending e-mails seems like a significant amount of work for something not many people would use).

[hwengmgr]

In this case, I have a HW problem that seems to be causing random events that make the hash rate look silly (GH/s). In reality nothing is actually happening. All I do is stop the miner and restart it (don't even have to exit the program) and things work again. So the fix is very simple. Maybe email is more than is needed. How about stop the miner if the hash rate changes dramatically, and just start it again? Setting some sort of flag on the screen that this happened would be good, or just put it into the log file.

I will start to look at the OCM code and see if I can figure out how to code this in.

[gertjaap]

It would really be helpful to catch the debug log with extended debug logging enabled (see #65). I don't think there's anything wrong and i suspect the miner is still running. It might just have switched from MH to kH for some reason and the software is not handling that properly. I don't think we should restart the miners on significant hashrate fluctuations. Could be other reasons (starting a game or something, occupying the GPU otherwise) that restarting won't fix.

[doudz]

Got the same problem, GeForce GT 710 reports 18 GH/s, I'm not sure this is real....

Trying to launch ccminer without OCM, I got "no kernel image is available for execution on the device"

[2020-09-29 13:26:40] GPU #0: GeForce GT 710, 18.51 GH/s
[2020-09-29 13:26:40] GPU #0: no kernel image is available for execution on the device
[2020-09-29 13:26:40] GPU #0: no kernel image is available for execution on the device
[2020-09-29 13:26:40] GPU #0: no kernel image is available for execution on the device