Software distributed shared memory protocols that adapt between single writer and multiple writer
Master of Science
We present two software distributed shared memory protocols that dynamically adapt between a single writer (SW) and a multiple writer (MW) protocol based on the application's sharing patterns. The first protocol adapts based on write-write false sharing, the second based on a combination of write-write false sharing and write granularity. The adaptation is automatic. No user or compiler information is needed. We measured the performance of these protocols on a test suite of eight applications, covering a broad spectrum in terms of write-write false sharing and write granularity. The adaptive protocols match or exceed the performance of the best of MW and SW in seven out of the eight applications. Speedup improvements over SW range from a factor of 1.02 to 2.7, and over MW from 1.02 to 1.6. In addition, memory usage is reduced considerably compared to MW, in some cases making the memory overhead all but negligible.