paperbot · PL 论文追踪

全部论文 待读 收藏

P2 Compiler-driven FPGA virtualization with SYNERGY

编译器驱动的 FPGA 虚拟化:SYNERGY
ASPLOS Proceedings of the 26th ACM International Conference on Architectural Support for Programming Languages and Operating Systems2021匹配度 2/10 · 论文主题是编译器驱动的 FPGA 虚拟化,属于编译/体系结构方向,与用户关注的静态分析、LLM/Agent、浮点程序分析或贝叶斯程序分析均无直接关联;仅“编译与代码生成”标签存在边缘交集。引用 22 inbox
Joshua Landgraf, Tiffany Yang, Will Lin, Christopher J. Rossbach, Eric Schkufza
crossref: 2026-09-12

AI 速览

一句话:提出 Synergy,通过 Verilog 编译变换和运行时支持在现有 FPGA 上实现挂起/恢复、迁移与多路复用。

现代数据中心需要在虚拟化基础设施上按需使用 FPGA,但制造商的硬件虚拟化方案难以提供通用支持。本文提出编译器/运行时协同的 FPGA 虚拟化方案 Synergy:对 Verilog 程序进行编译变换,使生成的硬件能够在符合原程序语义的前提下,以低于时钟周期的粒度将控制权交还给软件。基于该能力,Synergy 可在现有 FPGA 上高效支持核心虚拟化原语,包括挂起/恢复、程序迁移以及空间/时间复用。作者在 Altera SoC 集群和 Amazon F1 的 Xilinx FPGA 上虚拟化多种未修改的 FPGA 工作负载;结果显示性能为未虚拟化版本的 3–4 倍以内,且 FPGA 资源占用仅小幅增加。

问题
FPGA 在云数据中心面临多租户、整合与工作负载迁移需求,但缺乏通用的虚拟化策略,而硬件方案难以奏效。
方法
Synergy 通过 Verilog 编译变换,使硬件可按原程序语义在子时钟周期粒度让出控制权,并由运行时支持挂起/恢复、迁移以及空间/时间复用。
评估
在 Altera SoC 集群与 Amazon F1 的 Xilinx FPGA 上运行无需修改的工作负载,性能为未虚拟化版本的 3–4 倍以内,FPGA 资源占用小幅增加。
原文摘要(Abstract)

FPGAs are increasingly common in modern applications, and cloud providers now support on-demand FPGA acceleration in data centers. Applications in data centers run on virtual infrastructure, where consolidation, multi-tenancy, and workload migration enable economies of scale that are fundamental to the provider’s business. However, a general strategy for virtualizing FPGAs has yet to emerge. While manufacturers struggle with hardware-based approaches, we propose a compiler/runtime-based solution called Synergy. We show a compiler transformation for Verilog programs that produces code able to yield control to software at sub-clock-tick granularity according to the semantics of the original program. Synergy uses this property to efficiently support core virtualization primitives: suspend and resume, program migration, and spatial/temporal multiplexing, on hardware which is available today. We use Synergy to virtualize FPGA workloads across a cluster of Altera SoCs and Xilinx FPGAs on Amazon F1. The workloads require no modification, run within 3−4× of unvirtualized performance, and incur a modest increase in FPGA fabric utilization.

链接与引用

DOI 原文 · PDF(开放获取) ·

BibTeX
@inproceedings{paperbot6139,
  title = {Compiler-driven FPGA virtualization with SYNERGY},
  author = {Joshua Landgraf and Tiffany Yang and Will Lin and Christopher J. Rossbach and Eric Schkufza},
  booktitle = {ASPLOS},
  year = {2021},
  doi = {10.1145/3445814.3446755}
}