Design and Evaluation of SmallFloat SIMD extensions to the RISC-V ISA

被引:0
|
作者
Tagliavini, Giuseppe [1 ]
Mach, Stefan [2 ]
Rossi, Davide [1 ]
Marongiu, Andrea [3 ]
Benini, Luca [1 ,2 ]
机构
[1] Univ Bologna, DEI, Bologna, Italy
[2] Swiss Fed Inst Technol, IIS, Zurich, Switzerland
[3] Univ Bologna, DISI, Bologna, Italy
基金
欧盟地平线“2020”;
关键词
D O I
10.23919/date.2019.8714897
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
RISC-V is an open-source instruction set architecture (ISA) with a modular design consisting of a mandatory base part plus optional extensions. The RISC-V 32IMFC ISA configuration has been widely adopted for the design of new-generation, low-power processors. Motivated by the important energy savings that smaller-than-32-bit FP types have enabled in several application domains and related compute platforms, some recent studies have published encouraging early results for their adoption in RISC-V processors. In this paper we introduce a set of ISA extensions for RISC-V 32IMFC, supporting scalar and SIMD operations (fitting the 32-bit register size) for 8-bit and two 16-bit FP types. The proposed extensions are enabled by exposing the new FP types to the standard C/C++ type system and an implementation for the RISC-V GCC compiler is presented. As a further, novel contribution, we extensively characterize the performance and energy savings achievable with the proposed extensions. On average, experimental results show that their adoption provide benefits in terms of performance (1.64x speedup for 16-bit and 2.18x for 8-bit types) and energy consumption (30% saving for 16-bit and 50% for 8-bit types). We also illustrate an approach based on automatic precision tuning to make effective use of the new FP types.
引用
收藏
页码:654 / 657
页数:4
相关论文
共 50 条
  • [21] Pre-silicon evaluation and security enhancement for RISC-V AES extensions
    Zhao Y.
    Wei X.
    Li Y.
    He J.
    Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition), 2024, 52 (03): : 7 - 13
  • [22] A Custom Designed RISC-V ISA Compatible Processor for SoC
    Sharat, Kavya
    Bandishte, Sumeet
    Varghese, Kuruvilla
    Bharadwaj, Amrutur
    VLSI DESIGN AND TEST, 2017, 711 : 570 - 577
  • [23] Optimizing RISC-V ISA Usage by Sharing Coprocessors on MPSoC
    Lima, Pedro
    Vieira, Caio
    Reis, Jorge
    Almeida, Alexandre
    Silveira, Jarbas
    Goerl, Roger
    Marcon, Cesar
    21ST IEEE LATIN-AMERICAN TEST SYMPOSIUM (LATS 2020), 2020,
  • [24] RISC-V Instruction Set Architecture Extensions: A Survey
    Cui, Enfang
    Li, Tianzheng
    Wei, Qian
    IEEE ACCESS, 2023, 11 : 24696 - 24711
  • [25] FIXER: Flow Integrity Extensions for Embedded RISC-V
    De, Asmit
    Basu, Aditya
    Ghosh, Swaroop
    Jaeger, Trent
    2019 DESIGN, AUTOMATION & TEST IN EUROPE CONFERENCE & EXHIBITION (DATE), 2019, : 348 - 353
  • [26] DBPS: Dynamic Block Size and Precision Scaling for Efficient DNN Training Supported by RISC-V ISA Extensions
    Lee, Seunghyun
    Choi, Jeik
    Noh, Seockhwan
    Koo, Jahyun
    Kung, Jaeha
    2023 60TH ACM/IEEE DESIGN AUTOMATION CONFERENCE, DAC, 2023,
  • [27] Implementing the Draft RISC-V Scalar Cryptography Extensions
    Marshall, Ben
    Page, Daniel
    Pham, Thinh
    PROCEEDINGS OF THE 9TH INTERNATIONAL WORKSHOP ON HARDWARE AND ARCHITECTURAL SUPPORT FOR SECURITY AND PRIVACY, HASP 2020, 2020,
  • [28] Evaluating Cryptographic Extensions On A RISC-V Simulation Environment
    Sud, Parangat
    Neisarian, Shekoufeh
    Kavun, Elif Bilge
    2022 25TH EUROMICRO CONFERENCE ON DIGITAL SYSTEM DESIGN (DSD), 2022, : 548 - 555
  • [29] Simodense: a RISC-V softcore optimised for exploring custom SIMD instructions
    Papaphilippou, Philippos
    Kelly, Paul H. J.
    Luk, Wayne
    2021 31ST INTERNATIONAL CONFERENCE ON FIELD-PROGRAMMABLE LOGIC AND APPLICATIONS (FPL 2021), 2021, : 391 - 397
  • [30] Support Post Quantum Cryptography with SIMD Everywhere on RISC-V Architectures
    Wang, Liang-Ni
    Li, Ju-Hung
    Kuan, Chi-Bang
    Su, Yi-Chiao
    53RD INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING, ICPP 2024, 2024, : 23 - 32