A Resilient Interface for Approximate Data Access

被引:1
作者
Fabricio Filho, Joao [1 ,2 ]
Felzmann, Isaias B. [2 ]
Azevedo, Rodolfo J. [2 ]
Wanner, Lucas F. [2 ]
机构
[1] Univ Tecnol Fed Parana, Campus Campo Mourao, Campo Mourao, Brazil
[2] Univ Estadual Campinas, Inst Comp, Campinas, Brazil
来源
2019 IX BRAZILIAN SYMPOSIUM ON COMPUTING SYSTEMS ENGINEERING (SBESC) | 2019年
基金
巴西圣保罗研究基金会;
关键词
Approximate Computing; Approximate Memory; Hybrid Approximate SRAM; Near Threshold Voltage;
D O I
10.1109/sbesc49506.2019.9046069
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Approximate memories offer lower energy cost while introducing errors to applications. If such errors affect essential parts of the application, the execution may fail, decreasing outputs quality and energy savings. We present AxRAM, a memory architecture interface for approximate data that allows applications to benefit from energy savings provided by approximate memories, while improving quality of results and failure rates. AxRAM places critical application data into non-approximate memory regions and implements a resilient addressing scheme that reduces invalid data accesses that lead to application crashes. On a voltage-overscaled SRAM, AxRAM reduces 50.97% of application crashes on average when compared to an unprotected approximate memory. Compared to a system using non approximate SRAM, our proposal reduces energy consumption by 4536% on average while providing at least 80% output quality.
引用
收藏
页数:8
相关论文
共 19 条
  • [1] Is "Good Enough" Computing Good Enough?
    不详
    [J]. COMMUNICATIONS OF THE ACM, 2015, 58 (05) : 12 - 14
  • [2] Exact global histogram specification optimized for structural similarity
    Avanaki, Alireza Nasiri
    [J]. OPTICAL REVIEW, 2009, 16 (06) : 613 - 621
  • [3] Near-Threshold Voltage (NTV) Computing
    De, Vivek
    Vangal, Sriram
    Krishnamurthy, Ram
    [J]. IEEE DESIGN & TEST, 2017, 34 (02) : 24 - 30
  • [4] DARK SILICON AND THE END OF MULTICORE SCALING
    Esmaeilzadeh, Hadi
    Blem, Emily
    St Amant, Renee
    Sankaralingam, Karthikeyan
    Burger, Doug
    [J]. IEEE MICRO, 2012, 32 (03) : 122 - 134
  • [5] ADeLe: A description language for approximate hardware
    Felzmann, Isaias B.
    Susin, Matheus M.
    Duenha, Liana
    Azevedo, Rodolfo
    Wanner, Lucas F.
    [J]. FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE, 2020, 102 (102): : 245 - 258
  • [6] Impact of memory approximation on energy efficiency
    Felzmann, Isaias B.
    Fabricio Filho, Joao
    Azevedo, Rodolfo
    Wanner, Lucas F.
    [J]. 2018 SYMPOSIUM ON HIGH PERFORMANCE COMPUTING SYSTEMS (WSCAD 2018), 2018, : 53 - 60
  • [7] Fursin G., 2008, COLLECTIVE BENCHMARK
  • [8] Measuring the Impact of Memory Errors on Application Performance
    Gottscho, Mark
    Shoaib, Mohammed
    Govindan, Sriram
    Sharma, Bikash
    Wang, Di
    Gupta, Puneet
    [J]. IEEE COMPUTER ARCHITECTURE LETTERS, 2017, 16 (01) : 51 - 55
  • [9] Gouy Isaac, 2004, The Computer Language Benchmarks Game
  • [10] MiBench: A free, commercially representative embedded benchmark suite
    Guthaus, MR
    Ringenberg, JS
    Ernst, D
    Austin, TM
    Mudge, T
    Brown, RB
    [J]. WWC-4: IEEE INTERNATIONAL WORKSHOP ON WORKLOAD CHARACTERIZATION, 2001, : 3 - 14