Cores won't draw much while blocking on a memory access. If the core count vs core utilization trade-off is a tie, the napkin math expectation would be same thr

Cores won't draw much while blocking on a memory access. If the core count vs core utilization trade-off is a tie, the napkin math expectation would be same throughput, same power draw of cores, less power draw of cache for less cache.