If the harvester is loaded close to GPU capacity, there are problems with lookups when harvester starts, as it is trying to do lookups for the past 5 challenges. This basically chokes the GPU for quite a bit. It also causes a bunch of stales / long lookups on the pool side and that messes up charts (e.g., FoxyPool).
Would it be possible to add to config.yaml harvester entry (e.g., harvester: / past_lookups: False) to suppress this behavior?
If the harvester is loaded close to GPU capacity, there are problems with lookups when harvester starts, as it is trying to do lookups for the past 5 challenges. This basically chokes the GPU for quite a bit. It also causes a bunch of stales / long lookups on the pool side and that messes up charts (e.g., FoxyPool).
Would it be possible to add to config.yaml harvester entry (e.g., harvester: / past_lookups: False) to suppress this behavior?