An assessment of the MUSS operating system

Abstract

This paper presents a critical appraisal of the MUSS portable operating system. Development costs for the system are discussed, including the cost of transferring it to a new machine. Its performance is also examined, together with a description of the techniques used to assess performance.

Threaded Runtime Support for Execution of Fine Grain Parallel Code on Coarse Grain Multiprocessors

“…Kernel IPC must maintain consistency of data when message passing in a shared memory address space. This is addressed in [23] by keeping the sender from writing to the sent data, which resembles our approach. Similar work is found in the scheduling mechanism used for lightweight RPC in [3].…”

“…Essential to the approach described in Section 3.2.2 is the notion that the sender of a message should not be capable of changing it until the receiver is finished. An operating system called MUSS [23] utilized a technique similar to this in managing IPC. In MUSS, the sender cannot access the sent data again, eliminating any potential inconsistencies.…”