sum (map sum xs) == sum (concat xs)
The JVM can do similar optimizations because it knows the class hierarchy at runtime and therefore declares every method as final until it finds a class with a method that overrides the previous definition. It's also capable inlining dynamic dispatch by profiling the type of the object it's dispatching on. If 99% of all calls go to class X then it can inline the method of class X and catch the remaining 1% with a fallback clause. Of course this comes at a cost. You need to retain the bytecode, class metadata, profiling information and duplicated code that is generated from inlining in memory.
Applying an operation to a total workload is identical to:
1) Split the total workload into smaller workloads.
2) Apply the operation to each smaller workload.
3) Apply the operation to the results from step 2.
Each smaller workload in step 2 can be run in parallel.
While that may seem obvious for adding numbers or money it may be less obvious when you have different operations for example a method in an OO class hierarchy.