Managing SMT Resource Usage through Speculative Instruction Window Weighting

Hans Vandierendonck, André SeznecSPIRE 2009

Abstract

Simultaneous multithreading processors dynamically share processor resources between multiple\nthreads. In general, shared SMT resources may be managed explicitly, e.g. by dynamically setting queue occupation\nbounds for each thread as in the DCRA and Hill-Climbing policies. Alternatively, resources may be managed\nimplicitly, i.e. resource usage is controlled by placing the desired instruction mix in the resources. In this case,\nthe main resource management tool is the instruction fetch policy which must predict the behavior of each thread\n(branch mispredictions, long-latency loads, etc.) as it fetches instructions.\n\nIn this paper, we present the use of Speculative InstructionWindowWeighting (SIWW) to bridge the gap between\nimplicit and explicit SMT fetch policies. SIWW estimates for each thread the amount of outstanding work in the\nprocessor pipeline. Fetch proceeds for the thread with the least amount of work left. SIWW policies are implicit\nas fetch proceeds for the thread with the least amount of work left. They are also explicit as maximum resource allocation\ncan also be set. SIWWcan use and combine virtually any of the indicators that were previously proposed\nfor guiding the instruction fetch policy (number of in-flight instructions, number of low confidence branches,\nnumber of predicted cache misses, etc.). Therefore, SIWW is an approach to designing SMT fetch policies, rather\nthan a particular fetch policy.\n\nTargeting fairness or throughput is often contradictory and a SMT scheduling policy often optimizes only\none performance metric at the sacrifice of the other metric. Our simulations show that the SIWW fetch policy\ncan achieve at the same time state-of-the-art throughput, state-of-the-art fairness and state-of-the-art harmonic\nperformance mean.
BibTeX
@inproceedings{vandierendonck2009managing,
  author = {Hans Vandierendonck and André Seznec},
  title = {{Managing SMT Resource Usage through Speculative Instruction Window Weighting}},
  booktitle = {SPIRE},
  year = {2009},
}

Loading related content...

Figure