Improves G1 worst-case latencies by making the full GC parallel. The G1 garbage collector is designed to avoid full collections, but when the concurrent collections can't reclaim memory fast enough a fall back full GC will occur. The old implementation of the full GC for G1 used a single threaded mark-sweep-compact algorithm. With JEP 307 the full GC has been parallelized and now use the same amount of parallel worker threads as the young and mixed collections.