支持FPGA动态重构的RISC-V扩展指令集设计与实现

doi:10.19678/j.issn.1000-3428.0069267

计算机工程 ›› 2025, Vol. 51 ›› Issue (5): 229-238. doi: 10.19678/j.issn.1000-3428.0069267

支持FPGA动态重构的RISC-V扩展指令集设计与实现

周炫锦^1,2, 蔡刚^1,2, 黄志洪¹

1. 中国科学院空天信息创新研究院, 北京 100094;
2. 中国科学院大学电子电气与通信工程学院, 北京 100049

收稿日期:2024-01-21 修回日期:2024-03-22 出版日期:2025-05-15 发布日期:2025-05-10
通讯作者: 蔡刚,E-mail:caigang@aircas.ac.cn E-mail:caigang@aircas.ac.cn
基金资助:
国家自然科学基金(61704173)。

Design and Implementation of RISC-V Extended Instruction Set Supporting FPGA Dynamic Reconfiguration

ZHOU Xuanjin^1,2, CAI Gang^1,2, HUANG Zhihong¹

1. Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China;
2. School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100049, China

Received:2024-01-21 Revised:2024-03-22 Online:2025-05-15 Published:2025-05-10

摘要/Abstract

摘要： 目前实现动态重构的常用方法是通过片上接口进行配置,一般采用现场可编程门阵列(FPGA)官方提供的动态重构控制知识产权(IP)核,并通过系统总线与处理器相连。这种方法会占用较多静态部分的逻辑资源,并且限制了片上接口的运行频率。针对这些问题,提出将FPGA抽象为大规模存储器的设计理念,构建DPRC动态重构控制指令集及配套应用程序编程接口(API),以优化逻辑资源占用量,消除缓冲延迟。指令集的实现以原有RV32IMC为基础,采用微指令序列控制片上接口部分,通过与数据通路紧密耦合来减少逻辑资源使用量,使用参数化多周期方案优化时序并确保通用性。实验结果表明,与传统方法相比,该系统中动态重构功能相关逻辑资源占用量减少84%,频率提高312%。相较于原有处理器,添加扩展指令集后处理器自身资源占用量仅增加5%,最差情况下扩展部分对时钟周期的影响小于0.2 ns,表明该动态重构控制方案具有低成本、高主频的特性。

关键词: RISC-V指令集, 扩展指令集, 动态重构, FPGA技术, 大规模存储器

Abstract: Currently, dynamic refactoring is implemented by configuring it through on-chip interfaces, usually using the dynamic refactoring control Intellectual Property (IP) core provided by the official Field Programmable Gate Array (FPGA), and connected to the processor through the system bus. This method consumes a significant amount of static logic resources and limits the frequency of the on-chip interfaces. To address these issues, this study proposes the design concept of abstracting an FPGA as large-scale memory. A Dynamic Partial ReConfiguration (DPRC) dynamic reconstruction control instruction set and supporting Application Programming Interface (API) are constructed to optimize the usage of logic resources and eliminate buffer latency. The instruction set is implemented based on the original RV32IMC, using microinstruction sequences to control the on-chip interface part. By tightly coupling with the data path, logic resource usage is reduced, and a parameterized multicycle scheme is used to optimize the timing and ensure generality. Experimental results show that, compared with traditional methods, the dynamic reconstruction function in this system reduces the logical resource usage by 84% and increases the frequency by 312%. Compared with the original processor, the addition of an extended instruction set increases the resource usage of the processor alone by 5%. In the worst case, the impact of the extended part on the clock cycle is less than 0.2 ns, which shows that this dynamic reconstruction control scheme has the characteristics of low cost and high clock frequency.

Key words: RISC-V instruction set, extended instruction set, dynamic reconfiguration, FPGA technology, large scale memory

中图分类号:

TP332

周炫锦, 蔡刚, 黄志洪. 支持FPGA动态重构的RISC-V扩展指令集设计与实现[J]. 计算机工程, 2025, 51(5): 229-238.

ZHOU Xuanjin, CAI Gang, HUANG Zhihong. Design and Implementation of RISC-V Extended Instruction Set Supporting FPGA Dynamic Reconfiguration[J]. Computer Engineering, 2025, 51(5): 229-238.

https://www.ecice06.com/CN/Y2025/V51/I5/229

参考文献

[1] 胡明志,尹震宇,王军,等.国产FPGA通用以太网控制器的设计与实现[J].小型微型计算机系统, 2024, 45(10):2529-2536 HU M Z, YIN Z Y, WANG J, et al. Design and implementation of domestic FPGA universal Ethernet controller[J]. Journal of Chinese Computer Systems, 2024, 45(10):2529-2536.(in Chinese)
[2] CHUNG E, FOWERS J, OVTCHAROV K, et al. Serving DNNs in real time at datacenter scale with project brainwave[J]. IEEE Micro, 2018, 38(2):8-20.
[3] CAULFIELD A M, CHUNG E S, PUTNAM A, et al. A cloud-scale acceleration architecture[C]//Proceedings of the 49th Annual IEEE/ACM International Symposium on Microarchitecture. Washington D.C.,USA:IEEE Press,2016:1-13.
[4] SHEPOVALOV M, AKELLA V. FPGA and GPU-based acceleration of ML workloads on Amazon cloud-a case study using gradient boosted decision tree library[J]. Integration, 2020, 70:1-9.
[5] FPGA云服务器-百度智能云[EB/OL].[2023-05-22] . https://cloud.baidu.com/product/fpga.html. FPGA cloud server-Baidu cloud[EB/OL].[2023-05-22] . https://cloud.baidu.com/product/fpga.html.(in Chinese)
[6] FPGA加速云服务器(FPGA accelerated cloud server)_FACS_FPGA加速_华为云[EB/OL].[2023-05-22] . https://www.huaweicloud.com/product/facs.html. FPGA accelerated cloud server_FACS_FPGA_Huawei cloud[EB/OL].[2023-05-22] . https://www.huaweicloud.com/product/facs.html.(in Chinese)
[7] FPGA云服务器[EB/OL].[2023-05-22] . https://www.aliyun.com/product/ecs/fpga. FPGA cloud server[EB/OL].[2023-05-22] . https://www.aliyun.com/product/ecs/fpga.(in Chinese)
[8] DIAMANTOPOULOS D, POLIG R, RINGLEIN B, et al. Acceleration-as-a-service:a cloud-native Monte-Carlo option pricing engine on CPUs, GPUs and disaggregated FPGAs[C]//Proceedings of the 14th IEEE International Conference on Cloud Computing. Washington D.C.,USA:IEEE Press,2021:726-729.
[9] MENCER O, ALLISON D, BLATT E, et al. The history, status, and future of FPGAs:hitting a nerve with field-programmable gate arrays[J]. Queue, 2020, 18(3):71-82.
[10] MANEV K, POWELL J, MATAS K, et al. byteman:a bitstream manipulation framework[C]//Proceedings of the International Conference on Field-Programmable Technology. Washington D.C.,USA:IEEE Press,2022:1-9.
[11] HOFMANN M, TANG Z Y, ORGILL J, et al. XBERT:Xilinx logical-level bitstream embedded RAM transfusion[C]//Proceedings of the 29th IEEE Annual International Symposium on Field-Programmable Custom Computing Machines. Washington D.C.,USA:IEEE Press,2021:1-9.
[12] REVIRIEGO P, ULLAH A, PONTARELLI S. PR-TCAM:efficient TCAM emulation on Xilinx FPGAs using partial reconfiguration[J]. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2019, 27(8):1952-1956.
[13] CHUNG E S, HOE J C, MAI K. CoRAM:an in-fabric memory architecture for FPGA-based computing[C]//Proceedings of the 19th ACM/SIGDA International Symposium on Field Programmable Gate Arrays. New York,USA:ACM Press,2011:97-106.
[14] DAO N, ATTWOOD A, HEALY B, et al. FlexBex:a RISC-V with a reconfigurable instruction extension[C]//Proceedings of the International Conference on Field-Programmable Technology. Washington D.C.,USA:IEEE Press,2020:190-195.
[15] CARDONA L A, FERRER C. AC_ICAP:a flexible high speed ICAP controller[J]. International Journal of Reconfigurable Computing, 2015, 2015:9.
[16] WANG G H, LUO D M, WANG F Z, et al. A tiny and multifunctional ICAP controller for dynamic partial reconfiguration system[C]//Proceedings of the NASA/ESA Conference on Adaptive Hardware and Systems. Washington D.C.,USA:IEEE Press,2017:71-76.
[17] VIPIN K, FAHMY S A. ZyCAP:efficient partial reconfiguration management on the Xilinx Zynq[J]. IEEE Embedded Systems Letters, 2014, 6(3):41-44.
[18] SULTANA B, ULLAH A, ALI MALIK A, et al. VR-ZYCAP:a versatile resourse-level ICAP controller for ZYNQ SOC[J]. Electronics, 2021, 10(8):899.
[19] VIPIN K, FAHMY S A. FPGA dynamic and partial reconfiguration:a survey of architectures, methods, and applications[J]. ACM Computing Surveys, 2018, 51(4):1-39.
[20] WEERASINGHE J, POLIG R, ABEL F, et al. Network-attached FPGAs for data center applications[C]//Proceedings of 2016 International Conference on Field-Programmable Technology. Washington D.C.,USA:IEEE Press,2016:36-43.
[21] WEERASINGHE J, ABEL F, HAGLEITNER C, et al. Enabling FPGAs in hyperscale data centers[C]//Proceedings of 2015 IEEE International Conference on Ubiquitous Intelligence and Computing. Washington D.C.,USA:IEEE Press,2015:1078-1086.
[22] JADHAV A, MATTE P N, CHIPPALKATTI P. Configuration of FPGA through internal configuration access port[C]//Proceedings of the 2nd International Conference on Computing Methodologies and Communication. Washington D.C.,USA:IEEE Press,2018:497-499.
[23] PAPADIMITRIOU K, DOLLAS A, HAUCK S. Performance of partial reconfiguration in FPGA systems:a survey and a cost model[J]. ACM Transactions on Reconfigurable Technology and Systems, 2011, 4(4):1-24.
[24] HANSEN S G, KOCH D, TORRESEN J, et al. High speed partial run-time reconfiguration using enhanced ICAP hard macro[C]//Proceedings of the 2011 IEEE International Symposium on Parallel and Distributed Processing. Washington D.C.,USA:IEEE Press,2011:174-180.
[25] 王凯帆,徐易难,余子濠,等.香山开源高性能RISC-V处理器设计与实现[J].计算机研究与发展, 2023, 60(3):476-493. WANG K F, XU Y N, YU Z H, et al. XiangShan open-source high performance RISC-V processor design and implementation[J]. Journal of Computer Research and Development, 2023, 60(3):476-493.(in Chinese)
[26] ZHAO J, KORPAN B, GONZALEZ A, et al. SonicBOOM:The 3rd generation Berkeley out-of-order machine[EB/OL].[2023-08-05]. https://people.eecs.berkeley.edu/~krste/papers/SonicBOOM-CARRV2020.pdf#:~:text=SonicBOOM% 20is% 20an% 20open-source% 20RTL% 20implementation% 20of% 20a,core% 20by% 20IPC% 20available% 20at% 20time% 20of% 20publication.
[27] RISC-V microcontroller[EB/OL].[2023-08-05]. https://github.com/riscv-mcu/e203_hbirdv2.
[28] ASANOVI AĆG K, AVIZIENIS R, BACHRACH J, et al. The rocket chip generator:UCB/EECS-2016-17[EB/OL].[2023-08-05]. http://www2.eecs.berkeley.edu/Pubs/TechRpts/2016/EECS-2016-17.html.
[29] SpinalHDL[EB/OL].[2023-08-05]. https://github.com/SpinalHDL/VexRiscv.

选择文件类型/文献管理软件名称

选择包含的内容

支持FPGA动态重构的RISC-V扩展指令集设计与实现

Design and Implementation of RISC-V Extended Instruction Set Supporting FPGA Dynamic Reconfiguration

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 10

编辑推荐

Metrics

本文评价

[1]	刘坤, 张鹏程, 金惠颖, 吉顺慧. 云网融合环境下组合服务的动态重构[J]. 计算机工程, 2025, 51(5): 206-218.
[2]	雷斗威, 何德彪, 罗敏, 彭聪. 基于AVX512的格密码高速并行实现[J]. 计算机工程, 2024, 50(2): 15-24.
[3]	廖汉松, 吴朝晖, 李斌. 基于RISC-V的卷积神经网络专用指令集处理器[J]. 计算机工程, 2021, 47(7): 196-204.
[4]	李雨倩, 焦继业, 刘有耀, 郝振和. 嵌入式RISC-V乱序执行处理器的研究与设计[J]. 计算机工程, 2021, 47(2): 261-267,284.
[5]	邓哲, 张伟功, 朱晓燕, 杜瑞. 动态可重构总线数据传输管理方法设计与实现[J]. 计算机工程, 2013, 39(1): 264-269.
[6]	蒋静, 潘振宽, 殷宪振, 夏佰??. 基于最大应用模板的服务组合模型[J]. 计算机工程, 2010, 36(12): 69-71.
[7]	蒋静;夏佰强;潘振宽;殷宪振. 基于SOA的动态服务集成与调度网格模型[J]. 计算机工程, 2009, 35(22): 68-70.
[8]	马连博;张磊;胡琨元;苏卫星;金守义. 基于RFID的基带通信IP核的设计与仿真[J]. 计算机工程, 2009, 35(15): 275-276,.
[9]	张天序; ;郑伟; ;颜露新;. 图像处理系统中的新型动态重构技术[J]. 计算机工程, 2007, 33(17): 222-224.
[10]	孙中轶;何牧;蔡鸿明;姜丽红. 应用AOP设计的可动态重构工作流引擎架构[J]. 计算机工程, 2007, 33(06): 62-64.

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

支持FPGA动态重构的RISC-V扩展指令集设计与实现

Design and Implementation of RISC-V Extended Instruction Set Supporting FPGA Dynamic Reconfiguration

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 10

编辑推荐

Metrics

本文评价