一句话：识别虚拟化与内存分配器交互导致的地址转换瓶颈，提出细粒度物理内存预留 PTEMagnet 来加速嵌套页表遍历。
云计算在数据密集型任务中快速普及，应用常运行于虚拟化环境中，并与其他应用共置于同一虚拟机。该工作指出了一种新的地址转换瓶颈：虚拟化、共置与 Linux 内存分配器相互作用造成内存碎片，使 host 页表的有效缓存占用大于 guest 页表。膨胀的 host 页表缓存占用导致嵌套页表遍历期间频繁发生缓存缺失，从而增加页表遍历延迟。为缓解这一问题，作者提出 PTEMagnet，通过细粒度物理内存预留来减少碎片并压缩 host 页表的有效缓存占用，目标是加速公有云中的嵌套页表遍历。摘要未给出具体评估结果与量化数据。
The last few years have seen a rapid adoption of cloud computing for data-intensive tasks. In the cloud environment, it is common for applications to run under virtualization and to share a virtual machine with other applications (e.g., in a virtual private cloud setup). In this setting, our work identifies a new address translation bottleneck caused by memory fragmentation stemming from the interaction of virtualization, colocation, and the Linux memory allocator. The fragmentation results in the effective cache footprint of the host PT being larger than that of the guest PT. The bloated footprint of the host PT leads to frequent cache misses during nested page walks, increasing page walk latency.
@inproceedings{paperbot6156,
title = {PTEMagnet: fine-grained physical memory reservation for faster page walks in public clouds},
author = {Artemiy Margaritov and Dmitrii Ustiugov and Amna Shahab and Boris Grot},
booktitle = {ASPLOS},
year = {2021},
doi = {10.1145/3445814.3446704}
}