The Laplace Microarchitecture for Tracking Data Uncertainty and Its Implementation in a RISC-V Processor

被引:0
|
作者
Tsoutsouras, Vasileios [1 ,2 ]
Kaparounakis, Orestis [1 ]
Bilgin, Bilgesu [2 ]
Samarakoon, Chatura [2 ]
Meech, James [2 ]
Heck, Jan [2 ]
Stanley-Marbell, Phillip [1 ,2 ]
机构
[1] Signaloid, Cambridge, England
[2] Univ Cambridge, Cambridge, England
来源
PROCEEDINGS OF 54TH ANNUAL IEEE/ACM INTERNATIONAL SYMPOSIUM ON MICROARCHITECTURE, MICRO 2021 | 2021年
基金
英国工程与自然科学研究理事会;
关键词
uncertainty tracking; distributional representations; arithmetic on distributions; RISC-V; FUZZY; ARCHITECTURE;
D O I
10.1145/3466752.3480131
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
We present Laplace, a microarchitecture for tracking machine representations of probability distributions paired with architectural state. We present two newmethods for in-processor distribution representations which are approximations of probability distributions just as floating-point number representations are approximations of real-valued numbers. Laplace executes unmodified RISC-V binaries and can track uncertainty through them. We present two sets of ISA extensions to provide a mechanism to initialize distributional information in the microarchitecture and to allow applications to query statistics of the distributional information without exposing the uncertainty representations above the ISA. We evaluate the accuracy and performance of Laplace using a suite of 21 benchmarks spanning domains ranging from variational quantum algorithms and sensor data processing to materials properties modeling. Monte Carlo simulation on the benchmarks requires 2 076x more instructions on average (and up to 21 343x in some cases) to achieve the same accuracy that Laplace can achieve in a single execution. Compared to state-of-the-art alternatives to Monte Carlo, Laplace achieves an average 1.3x accuracy improvement versus PaCAL [22] and more than 4.6x accuracy improvement versus the method used by the NIST Uncertainty Machine [26], quantified using the Wasserstein distance to Monte Carlo. Unlike existing methods for uncertainty tracking which require software to be rewritten in a domain-specific language or extensive source-level changes, Laplace achieves all of these benefits while requiring no changes to existing binaries in order to track uncertainty through them, with only minimal changes required to get uncertainty information into the microarchitecture. We have deployed an implementation of Laplace as a commercial product in the form of a cloud-accessible virtual machine.
引用
收藏
页码:1254 / 1269
页数:16
相关论文
共 50 条
  • [21] A RISC-V ISA Compatible Processor IP for SoC
    Budi, Suseela
    Gupta, Pradeep
    Varghese, Kuruvilla
    Bharadwaj, Amrutur
    2018 INTERNATIONAL SYMPOSIUM ON DEVICES, CIRCUITS AND SYSTEMS (ISDCS), 2018,
  • [22] Design and Verification Environment for RISC-V Processor Cores
    Oleksiak, Adrian
    Cieslak, Sebastian
    Marcinek, Krzysztof
    Pleskacz, Witold A.
    PROCEEDINGS OF THE 2019 26TH INTERNATIONAL CONFERENCE MIXED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS (MIXDES 2019), 2019, : 206 - 209
  • [23] Parallel Implementation of PIPO Block Cipher on 32-bit RISC-V Processor
    Kwak, YuJin
    Kim, YoungBeom
    Seo, Seog Chung
    INFORMATION SECURITY APPLICATIONS, 2021, 13009 : 183 - 193
  • [24] XiangShan Open-Source High Performance RISC-V Processor Design and Implementation
    Wang K.
    Xu Y.
    Yu Z.
    Tang D.
    Chen G.
    Chen X.
    Gou L.
    Hu X.
    Jin Y.
    Li Q.
    Li X.
    Lin J.
    Liu T.
    Liu Z.
    Wang H.
    Wang H.
    Zhang C.
    Zhang F.
    Zhang L.
    Zhang Z.
    Zhang Z.
    Zhao Y.
    Zhou Y.
    Zou J.
    Cai Y.
    Huan D.
    Li Z.
    Zhao J.
    He W.
    Sun N.
    Bao Y.
    Jisuanji Yanjiu yu Fazhan/Computer Research and Development, 2023, 60 (03): : 476 - 493
  • [25] A Minimal RISC-V Vector Processor for Embedded Systems
    Johns, Matthew
    Kazmierski, Tom J.
    PROCEEDINGS OF THE 2020 FORUM FOR SPECIFICATION AND DESIGN LANGUAGES (FDL), 2020,
  • [26] Efficient resource shared RISC-V multicore processor
    Islam, Md Ashraful
    Kise, Kenji
    2021 IEEE 14TH INTERNATIONAL SYMPOSIUM ON EMBEDDED MULTICORE/MANY-CORE SYSTEMS-ON-CHIP (MCSOC 2021), 2021, : 366 - 372
  • [27] Integrated Dynamic Memory Manager for a RISC-V Processor
    Tsai, Chun-Jen
    Chao, Chun Wei
    Hong, Sheng-Di
    2023 IFIP/IEEE 31ST INTERNATIONAL CONFERENCE ON VERY LARGE SCALE INTEGRATION, VLSI-SOC, 2023, : 277 - 281
  • [28] Single Cycle RISC-V Micro Architecture Processor and its FPGA Prototype
    Dennis, Don Kurian
    Priyam, Ayushi
    Virk, Sukhpreet Singh
    Agrawal, Sajal
    Sharma, Tanuj
    Mondal, Arijit
    Ray, Kailash Chandra
    2017 7TH INTERNATIONAL SYMPOSIUM ON EMBEDDED COMPUTING AND SYSTEM DESIGN (ISED), 2017,
  • [29] Modeling RISC-V processor in IP-XACT
    Pekkarinen, Esko
    Hamalainen, Timo D.
    2018 21ST EUROMICRO CONFERENCE ON DIGITAL SYSTEM DESIGN (DSD 2018), 2018, : 140 - 147
  • [30] Reconfigurable RISC-V Secure Processor And SoC Integration
    Zang, Zhenya
    Liu, Yao
    Cheung, Ray C. C.
    2019 IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL TECHNOLOGY (ICIT), 2019, : 827 - 832