1993], and compare the resulting code with the code obtained from an aggressive graph coloring algorithm that performs iterated register coalescing [George and Appel 1996].
Nonetheless, the resulting code is quite efficient: on the benchmarks we studied, it is within 12% as fast as code generated by an aggressive graph coloring algorithm for all but two benchmarks.
For each of these benchmarks, we compare linear scan register allocation against (1) a well-tuned graph coloring algorithm and (2) a simple "usage count" register allocation scheme.
As before, we compare linear scan register allocation against a graph coloring algorithm and the simple algorithm based on usage counts.
The figure indicates that linear scan allocation (L) can be considerably faster than even a simple and fast graph coloring algorithm (C).
Processor graph coloring algorithms. Several parts of the parallel AMG algorithm require the formation of a graph incidence matrix of size p on a processor.
For example, parallel graph coloring algorithms [8] provide a significantly better probabilistic performance bound than the deterministic algorithms described here.