Making User-Level VMM for Deterministic Parallelism Nonblocking and Efficient | IEEE Conference Publication | IEEE Xplore