Description of the bug in the issues.
The miner/worker should
- Run one SDR per core group until all groups are used
- Additional SDR per group if
(1 + FIL_PROOFS_MULTICORE_SDR_PRODUCERS) * [SDR count] <= [core count in group]
(dedicated FIL_PROOFS_MULTICORE_SDR_PRODUCERS) - loop step 2
Other strategies seem possible:
- Run one SDR per core group until all groups are used
- Additional SDR per group if
([core count in group] - [SDR count] - FIL_PROOFS_MULTICORE_SDR_PRODUCERS) >= 0
(shared FIL_PROOFS_MULTICORE_SDR_PRODUCERS) - loop step 2
Maybe even a config switch to allow miners to choose what strategy to use.