Title :
Analytic models of workload behavior and pipeline performance
Author :
Squillante, Mark S. ; Kaeli, David R. ; Sinha, Himanshu
Author_Institution :
Res. Div., IBM Thomas J. Watson Res. Center, Yorktown Heights, NY, USA
Abstract :
The evaluation of pipeline performance and the analysis of different design alternatives and cost/performance tradeoffs are a fundamental aspect of high-performance computer system design. This performance evaluation process requires accurate models of both the pipeline organization and the characteristics of the workload being executed. We derive general mathematical models and analyses of workload behavior and pipeline performance that can provide measures as accurate as detailed trace-driven simulations with the computational efficiency of analytic methods
Keywords :
multiprocessing systems; performance evaluation; pipeline processing; virtual machines; accurate models; analytic methods; computational efficiency; cost/performance tradeoffs; detailed trace-driven simulations; general mathematical models; high-performance computer system design; multiple instructions; performance evaluation process; pipeline organization; pipeline performance; pipelined processors; workload behavior; Computational modeling; Costs; Design engineering; Frequency measurement; Hazards; High performance computing; Mathematical model; Performance analysis; Pipeline processing; System analysis and design;
Conference_Titel :
Modeling, Analysis, and Simulation of Computer and Telecommunication Systems, 1997. MASCOTS '97., Proceedings Fifth International Symposium on
Conference_Location :
Haifa
Print_ISBN :
0-8186-7758-9
DOI :
10.1109/MASCOT.1997.567589
