Summary: A Comparison of the
M­PCP, D­PCP, and FMLP on LITMUS RT
Bj˜orn B. Brandenburg and James H. Anderson
The University of North Carolina at Chapel Hill
Dept. of Computer Science
Chapel Hill, NC 27599­3175 USA
{bbb,anderson}@cs.unc.edu
Abstract. This paper presents a performance comparison of three mul­
tiprocessor real­time locking protocols: the multiprocessor priority ceil­
ing protocol (M­PCP), the distributed priority ceiling protocol (D­PCP),
and the flexible multiprocessor locking protocol (FMLP). In the FMLP,
blocking is implemented via either suspending or spinning, while in the
M­PCP and D­PCP, all blocking is by suspending. The presented com­
parison was conducted using a UNC­produced Linux extension called
LITMUS RT . In this comparison, schedulability experiments were con­
ducted in which runtime overheads as measured on LITMUS RT were
used. In these experiments, the spin­based FMLP variant always exhib­
ited the best performance, and the M­PCP and D­PCP almost always
exhibited poor performance. These results call into question the prac­
tical viability of the M­PCP and D­PCP, which have been the de­facto

Source: Anderson, James - Department of Computer Science, University of North Carolina at Chapel Hill

Collections: Computer Technologies and Information Sciences