MEGATRON: Software-Managed Device TLB for Shared-Memory FPGA Virtualization
Yanqiang Liu, Jiacheng Ma, Zhengjun Zhang, Linsheng Li, Zhengwei Qi, Haibing Guan
摘要
FPGAs are being virtualized to improve resource utilization in data centers. Memory access performance is essential to FPGA hypervisors for shared-memory FPGA platform, where accelerators access memory spontaneously. DMA remapping with IOMMU provides a handy solution; however, fixed IOMMU can not benefit from the reconfigurability of FPGAs. In this work, we propose MEGATRON, a hybrid address translation service consisting of a hardware TLB and a software page table walker. By integrating MEGATRON into an existing FPGA hypervisor, we conduct a comprehensive analysis of link performance of a multi-link CPU-FPGA platform, and demonstrate the competitiveness of the customizable translation service.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
引用它的顶会 Paper1
问问它们各自怎么用它相关 Paper
- HDPAT: Hierarchical Distributed Page Address Translation for Wafer-Scale GPUsDaoxuan Xu, Ying Li, Yuwei Sun, Jie Ren 等HPCA 2026 · 被引用 1 次
- A Hypervisor for Shared-Memory FPGA PlatformsJiacheng Ma, Gefei Zuo, Kevin Loughlin, Xiaohe Cheng 等ASPLOS 2020 · 被引用 64 次
- Heliostat: Harnessing Ray Tracing Accelerators for Page Table WalksYuan Feng, Yuke Li, Jiwon Lee, Won Woo Ro 等ISCA 2025 · 被引用 8 次
- Translation Pass-Through for Near-Native Paging Performance in VMsShai Bergman, Mark Silberstein, Takahiro Shinagawa, Peter R. Pietzuch 等USENIX ATC 2023 · 被引用 10 次
- HyperTRIO: Hyper-Tenant Translation of I/O AddressesAlexey Lavrov, David WentzlaffISCA 2020 · 被引用 9 次
