Title :
On the costs and benefits of stochasticity in stream processing
Author :
Nadakuditi, Raj R. ; Markov, Igor L.
Author_Institution :
Univ. of Michigan, Ann Arbor, MI, USA
Abstract :
With the end of clock-frequency scaling, parallelism has emerged as the key driver of chip-performance growth. Yet, several factors undermine efficient simultaneous use of on-chip resources, which continue scaling with Moore´s law. These factors are often due to sequential dependencies, as illustrated by Amdahl´s law. Quantifying achievable parallelism can help prevent futile programming efforts and guide innovation toward the most significant challenges. To complement Amdahl´s law, we focus on stream processing and quantify performance losses due to stochastic runtimes. Using spectral theory of random matrices, we derive new analytical results and validate them by numerical simulations. These results allow us to explore unique benefits of stochasticity and show that they outweigh the costs for software streams.
Keywords :
logic design; matrix algebra; stochastic processes; Amdahl law; Moore law; clock-frequency scaling; random matrices; spectral theory; stochasticity; stream processing; Algorithm design and analysis; Costs; Decoding; Delay; Numerical simulation; Parallel processing; Pipelines; Runtime; Stochastic processes; Streaming media; Stream computing; latency; stochasticity;
Conference_Titel :
Design Automation Conference (DAC), 2010 47th ACM/IEEE
Conference_Location :
Anaheim, CA
Print_ISBN :
978-1-4244-6677-1