I just had a realization, is the checking cost O(n) to find out if there are conflicts where n is the number of processes?
kk
I think n should be the number of concurrent transactions in this case instead of processes. Also, if every transaction accesses a lot of memory addresses, then the number of different memory addresses might factor in as well.
I just had a realization, is the checking cost O(n) to find out if there are conflicts where n is the number of processes?
I think n should be the number of concurrent transactions in this case instead of processes. Also, if every transaction accesses a lot of memory addresses, then the number of different memory addresses might factor in as well.