Summary: Improving the Performance of Software
Distributed Shared Memory with Speculation
Michael Kistler, Member, IEEE, and Lorenzo Alvisi, Member, IEEE
Abstract--We study the performance benefits of speculation in a release consistent software distributed shared memory system. We
propose a new protocol, Speculative Home-based Release Consistency (SHRC), that speculatively updates data at remote nodes to
reduce the latency of remote memory accesses. Our protocol employs a predictor that uses patterns in past accesses to shared
memory to predict future accesses. We have implemented our protocol in a release consistent software distributed shared memory
system that runs on commodity hardware. We evaluate our protocol implementation using eight software distributed shared memory
benchmarks and show that it can result in significant performance improvements.
Index Terms--Distributed shared memory, protocol design and analysis, speculation.
A distributed shared memory (DSM) system allows a
collection of computers (nodes), connected by a high-
speed network, to be used as a single computing resource.
Applications can use the familiar shared-memory program-
ming model, but still benefit from the additional processing
power available in the system. To provide the illusion of
shared memory, the DSM system intercepts accesses to data
that physically resides in a remote node and brings the data to