A cache memory prioritizing long waiting load instructions

被引:0
|
作者
Hasegawa, Yusuke [1 ]
Suzuki, Ken-ichi [1 ]
Nakamura, Tadao [1 ]
机构
[1] Tohoku Univ, GSIS, Sendai, Miyagi 9808579, Japan
关键词
memory hierarchy; cache memory; criticality;
D O I
暂无
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
In this paper, we propose a criticality-based cache memory system preventing long waiting load instructions from cache misses, based on the latency of load instructions to be issued. These load instructions have a high possibility that they are contained in a long dependence instruction sequence, resulting in the degradation of performance. Furthermore, as the latency of instructions waiting for their issue becomes long, these instructions occupy instruction windows for a long time. We consider that load instructions occupying instruction windows for a long time deteriorate the performance of computer systems. Such a long wait in the instruction window occurs because of deep dependence on other instructions or long memory accesses due to cache misses of other load instructions. By reducing cache misses caused by these critical load instructions, the increase of the critical path becomes preventable. Also, instruction windows can issue more instructions efficiently, and exploit more empty space to make use of instruction level parallelism, leading to the improvement of the performance of computer systems.
引用
收藏
页码:663 / 671
页数:9
相关论文
共 50 条
  • [1] METHOD FOR PRIORITIZING WAITING ARITHMETIC INSTRUCTIONS.
    DeGroot, R.D.
    IBM technical disclosure bulletin, 1984, 27 (7 A): : 3735 - 3736
  • [2] Waiting for Cache
    Walton, James
    WESTERLY, 2022, 67 (02): : 198 - 204
  • [3] Classifying load and store instructions for memory renaming
    Reinman, Glenn
    Calder, Brad
    Tullsen, Dean
    Tyson, Gary
    Austin, Todd
    Proceedings of the International Conference on Supercomputing, 1999, : 399 - 407
  • [4] A Machine Learning Methodology for Cache Memory Design Based on Dynamic Instructions
    Navarro, Osvaldo
    Yudi, Jones
    Hoffmann, Javier
    Hernandez, Hector Gerardo Munoz
    Huebner, Michael
    ACM TRANSACTIONS ON EMBEDDED COMPUTING SYSTEMS, 2020, 19 (02)
  • [5] Implementation of a novel cache memory unit for storing processed data and instructions
    Sarkar, Saurav
    2017 INTERNATIONAL CONFERENCE ON NEXTGEN ELECTRONIC TECHNOLOGIES: SILICON TO SOFTWARE (ICNETS2), 2017, : 6 - 9
  • [6] Prioritizing surgical waiting lists
    Testi, A.
    Tanfani, E.
    Valente, R.
    Ansaldo, G. L.
    Torre, G. C.
    JOURNAL OF EVALUATION IN CLINICAL PRACTICE, 2008, 14 (01) : 59 - 64
  • [7] Thorough Evaluation of GPU Shared Memory Load and Store Instructions
    Okamoto, Satoshi
    Ito, Yasuaki
    Nakano, Koji
    Bordim, Jacir L.
    2014 SECOND INTERNATIONAL SYMPOSIUM ON COMPUTING AND NETWORKING (CANDAR), 2014, : 614 - 616
  • [8] A FEASIBILITY STUDY ON A SEPARATE CACHE MEMORY SYSTEM DIVIDED INTO 2 PARTS FOR INSTRUCTIONS AND FOR DATA
    SAITO, M
    NEC RESEARCH & DEVELOPMENT, 1982, (66): : 51 - 58
  • [9] Prioritizing the Surgical Waiting List-Cosine Consistency Index: An Optimized Framework for Prioritizing Surgical Waiting List
    Petwal, Hemant
    Rani, Rinkle
    JOURNAL OF MEDICAL IMAGING AND HEALTH INFORMATICS, 2020, 10 (12) : 2876 - 2892
  • [10] Cache Memory Design With Magnetic Skyrmions in a Long Nanotrack
    Chen, Mei-Chin
    Ranjan, Ashish
    Raghunathan, Anand
    Roy, Kaushik
    IEEE TRANSACTIONS ON MAGNETICS, 2019, 55 (08)