Multiprocessor performance
著者
書誌事項
Multiprocessor performance
（Wiley series in parallel computing）
Wiley, c1989
大学図書館所蔵 件 / 全23件
-
該当する所蔵館はありません
- すべての絞り込み条件を解除する
注記
Includes bibliographies and index
内容説明・目次
内容説明
This book describes the different functions and applications of the multiprocessor, starting with its architecture, structure and machine performance. It also uses examples of specific multiprocessor models.
目次
- Part 1 Multiprocessor architecture, program structure and machine performance: performance potential and limitations of parallel computer architectures
- performance issues in some typical multiprocessor architectures
- different types of multiprocessors. Part 2 Multiprocessor speed-up and Amdahl's law: an amendment to Amdahl's law
- a model of intrinsic programme behaviour. Part 3 Performance of interconnection networks: bandwidth of the Baseling and Omega networks with the DROP approach
- numerical solution of the DROP model
- bandwidth of the DROP model
- the HOLD network. Part 4 Parallel programme performance - series-parallel programme structures: the SPTG model of programme structure
- distribution of programme execution or completion time
- numerical solution for the execution time distribution function. Part 5 General acyclic random graphs as models of parallel programmes: asymptotic properties of the speed-up
- asymptotic analysis of the mean processing time
- comparison of the approximation with simulation results
- maximum likelihood estimation of p the probability of precedence between tasks
- task graphs with communication times between tasks. Part 6 Multiprocessor performance with task-graph models: the multiprocessor system model
- derivation of the job response time
- numerical examples and model validation
- extension of the method to the analysis of multiprocessor systems with different types of parallel programmes
- the case of systems with processors of different types. Part 7 Supercomputer performance evaluation: performance of a single processor
- performance of supercomputers having multiple vector processors
- supercomputers with a limited number of vector processing facilities
- Los Alamos benchmark characteristics. Part 8 Performance analysis of the connection machine: ROUTER network performance
- highly balanced computations
- the connection machine with floating point processors.
「Nielsen BookData」 より