Simulation Analysis of Data Sharing in Shared Memory Multiprocessors

Abstract

This dissertation examines shared memory reference patterns in parallel programs that run on bus-based, shared memory multiprocessors. The study reveals two distinct modes of sharing behavior. In sequential sharing, a processor makes multiple, sequential writes to the words within a block, uninterrupted by accesses from other processors. Under fine-grain sharing, processors contend for these words, and the number of per-processor sequential writes is low. Whether a program exhibits sequential or fine-grain sharing affects several factors relating to multiprocessor performance: the accuracy of sharing models that predict cache coherency overhead, the cache miss ratio and bus utilization of parallel programs, and the choice of coherency protocol. An architecture-independent model of write sharing was developed, based on the inter-processor activity to write-shared data. The model was used to predict the relative coherency overhead of write-invalidate and write-broadcast protocols. Architecturally detailed simulations validated the model for write-broadcast. Successive refinements produced acceptable predictions for write-invalidate. Block size was crucial for modeling write-invalidate, because the pattern of memory references within a block determines protocol performance. The cache and bus behavior of parallel programs running under write-invalidate protocols was evaluated over various block and cache sizes. A cross-protocol comparison provided empirical evidence of the performance loss caused by increasing block size in write-invalidate protocols and cache size in write-broadcast It then measured the extent to which read broadcast improved write-invalidate performance and competitive snooping helped write-broadcast The results indicated that read-broadcast reduced the number of invalidation misses, but at a high cost in processor lockout from the cache. The surprising net effect was an increase in total execution cycles.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 24, 1989
Accession Number
ADA631679

Entities

People

  • Susan J. Eggers

Organizations

  • University of California, Berkeley

Tags

Communities of Interest

  • Ground and Sea Platforms
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Access Time
  • Accuracy
  • Computer Architecture
  • Computer Programming
  • Computer Science
  • Computer-Aided Design
  • Computers
  • Data Analysis
  • Debugging
  • Load Monitoring
  • Multiprocessors
  • Operating Systems
  • Parallel Computing
  • Parallel Processing
  • Parallel Processors
  • Simulations
  • Simulators

Fields of Study

  • Computer science

Readers

  • Computational Modeling and Simulation
  • Parallel and Distributed Computing.