How do the processors actually decide which segment of the tree to work on? Specifically,how exactly do the processors decide where the slice of the processor before ended and it should begin its slice? Don't we end up wasting the work used to traverse the tree because each processor does it once? Since this is a blocking step, I feel like more attention could be paid on how to parallelize this process, since there's only one true allocation of work on the whole tree and each processor is now calculating its own version of the cost zones just so it knows which cost zone to work in.
How do the processors actually decide which segment of the tree to work on? Specifically,how exactly do the processors decide where the slice of the processor before ended and it should begin its slice? Don't we end up wasting the work used to traverse the tree because each processor does it once? Since this is a blocking step, I feel like more attention could be paid on how to parallelize this process, since there's only one true allocation of work on the whole tree and each processor is now calculating its own version of the cost zones just so it knows which cost zone to work in.