Journal of KIISE:Computer Systems and Theory (한국정보과학회논문지:시스템및이론)

Korean Institute of Information Scientists and Engineers (한국정보과학회)
권호 표지

A Hybrid Value Predictor using Speculative Update in Superscalar Processors

슈퍼스칼라 프로세서에서 모험적 갱신을 사용한 하이브리드 결과값 예측기

  • 박홍준 (극동정보대학 전산정보처리과) ;
  • 신영호 ((주)에어미디어 기술연구소) ;
  • 조영일 (수원대학교 컴퓨터과학과)
  • Published : 2001.12.01
Download PDF
⟨ Previous Next ⟩

Abstract

To improve the performance of wide-issue Superscalar microprocessors, it is essential to increase the width of instruction fetch and issue rate. Data dependences are major hurdle to exploit ILP(Instruction-Level Parallelism) efficiently, so several related works have suggested that the limits imposed by data dependences can be overcome to some extent with the use of the data value prediction. But the suggested mechanisms may access the same value prediction table entry again before they have been updated with a real data value. They will cause incorrect value prediction by using stable data and incur misprediction penalty and lowering performance. In this paper, we propose a new hybrid value predictor which achieve high performance by reducing stale data. Because the proposed hybrid value predictor can update the prediction table speculatively, it efficiently reduces the number of mispredicted instruction due to stable due to stale data. For SPECint95 benchmark programs on the 16-issue superscalar processors, simulation results show that the average prediction accuracy increase from 59% for non-speculative update to 72% for speculative update.

슈퍼스칼라 프로세서는 성능향상을 위해 명령어 반입폭과 이슈율을 증가시키고 있다. 데이터 종속성은 ILP(Instruction-Level Parallelism)를 향상시키는데 주요 장애요소가 되고 있으며, 최근 여러 논문에서 데이터 종속성을 제거하기 위해서 명령어의 결과값을 예상하는 메커니즘이 연구되었다. 그러나 이러한 예측기들은 예상한 명령어의 실제 결과값으로 예상 테이블을 갱신하기 전에 그 명령어를 다시 예상할 때 부적절(stale)한 데이터를 사용함으로써 예상 실패율이 증가하여 프로세서의 성능을 감소시킨다. 본 논문에서는 부적절 데이터 사용을 줄여 높은 성능을 얻을 수 있는 새로운 하이브리드 예측 메커니즘을 제안한다. 제안된 하이브리드 결과값 예측기는 예상 테이블을 모험적으로 갱신할 수 있기 때문에 부적절 데이터로 인해 잘못 예상되는 명령어의 수를 효과적으로 감소시킨다. 16-이슈폭 슈퍼스칼라 프로세서에서 SPECint95 벤치마크 프로그램에 대해 모험적 갱신을 사용함으로써 모험적 갱신을 사용하지 않은 경우의 평균 예상 정확도 59%에 비해 평균 예상 정확도가 72%에 비해 평균 예상 정확도가 72%로 크게 향상되었다.

Keywords

References

  1. M. H. Lipasti, C. B. Wilderson, J. P. Shen, 'Value Locality and Load Vaule Prediction,' Proc. of the 7th Intl, Conference on Architectural Support for Programming Languages and Operating Systems, pp.138-147, October 1996
  2. M. H. Lipasti and J. P. Shen, 'Exceeding the Dataflow limit via Value Prediction,' Proc. of the 29th Intl. Symp. on Microarchitecture, pp.226-237, December 1996 https://doi.org/10.1109/MICRO.1996.566464
  3. Y. Sazeides and J. E. Smith, 'The Predictability of Data Values,' Proc. of the 29th Intl. Symp. on Microarchitecture, pp.226-237, December 1997 https://doi.org/10.1109/MICRO.1997.645815
  4. F. Gabbay and A. Mendelson, 'Can Program Profiling Support Value prediction?' Proc, of the 30th Intl, Symp. on Microarchitecture, pp.270-280, December 1997 https://doi.org/10.1109/MICRO.1997.645817
  5. T-Y Yeh and Y. N. Patt, 'Alternative Implementations of Two-Level Adaptive Branch Prediction,' Proc, of the 19th Intl. Symposium on Computer Architecture, pp.124-134, 1992 https://doi.org/10.1145/139669.139709
  6. K. Wang and M. Franklin, 'Highly Accurate Data Value Prediction using Hybrid Predictors,' Proc. of the 30th International Symp, On Microarchitecture, pp.281-290, December 1997 https://doi.org/10.1109/MICRO.1997.645819
  7. J. Gonzalez and A. Gonzalez, 'The potential of data value speculation to boost ilp,' in 12th International Conference on Supercomputing, 1998 https://doi.org/10.1145/277830.277840
  8. G. Reinman and B. Calder, 'Predictive techniques for aggressive load speculation,' Proc. of the 31th International Symp. on Microarchitecture, 1998 https://doi.org/10.1109/MICRO.1998.742775
  9. T. Nakra, R. Gupta and M.L. Soffa, 'Global Context-Based Value Prediction', Proc. of the 5th Intl. Symposium on High Performance Computer Architecture, January 1999 https://doi.org/10.1109/HPCA.1999.744311
  10. D.C. Burger and T.M. Austin. 'The simplescalar tool set, version 2.0', Technical Report CS-TH97-1342, University of wisconsin, Madison. June 1997
  11. F. Dahlgren, F. Dahlgren and P. Stenstrom, 'Evaluation of Hardware-Based Stride and Sequential Prefetching in Shared-Memory Multiprocessors,' IEEE Transactions on Parallel and Distributed Systems. vol. 7. no. 4. pp. 385-398. April 1996 https://doi.org/10.1109/71.494633
  12. B. Calder, G. Reinman, D M. Tullsen, 'Selective Value Prediction,' Proc, of the 26th Intl. Symposium on Computer Architecture, May 1999 https://doi.org/10.1109/ISCA.1999.765940
  13. S. J. Lee, P. C. Yew, 'On Some Implementation Issues for Value Prediction on Wide-Issue ILP Processors', Intl. Conference on Parallel Architectures and Computer Technique, 2000 https://doi.org/10.1109/PACT.2000.888339