Pause instead of abort when temperature limit is reached

[arex1337]

I don't have any problems except when I forget to use the --gpu-watchdog=105 option I still feel this is an issue that would be awesome if implemented as I'm sure many people have felt the same way as me