No filtering required. When a potentially solved block arrives, spawn a thread to start validating it. When another potentially solved block arrives, spawn another thread to start validating it. First one to validate is the one you build your candidate for the next round atop.
That could improve a bit if there few blocks like this. But even if you can work the issue of shared access & deps, then what if there 100 such blocks ? Or 1000 ?
It can maybe just improve the resilience if there is few of them, but it's just pushing the pb one thread away, and there isnt an infinity of processing power on a computer, even with threads.
Hence the real solution is more on how to avoid wasting processing on those blocks, rather attempting at processing them as long as no better block is validated. Instead of increasing this wasted time on several thread to buffer a bit the excessive processing time.