article · 2024
Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table
Sungbin Jang, Junhyeok Park, Osang Kwon, Yongho Lee, Seokin Hong
Cited by 5Open access
Abstract
GPU memory virtualization has become essential for efficient programming, memory management, and address space sharing among computing devices in heterogeneous systems. Conventional GPU virtual memory systems use multi-level Radix Page Tables (RPTs) to store virtual-to-physical address mapping in device (GPU) memory. When a TLB miss occurs, a page table walker accesses each level of the page table sequentially to find the desired mapping. These sequential accesses significantly degrade performance, adding pressure to the GPU memory hierarchy. To make matters worse, recent computing systems now support five-level RPTs, further increasing the number of memory accesses required per page table walk.
Cite this paper
Jang, S., Park, J., Kwon, O., Lee, Y., & Hong, S. (2024). Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table. https://doi.org/10.1145/3656019.3676900
Jang, Sungbin, et al. “Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table.” 2024, pp. 325–37, https://doi.org/10.1145/3656019.3676900.
Jang, Sungbin, Junhyeok Park, Osang Kwon, Yongho Lee, and Seokin Hong. 2024. “Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table.” https://doi.org/10.1145/3656019.3676900.
Jang, S. et al. (2024) ‘Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table’. Available at: https://doi.org/10.1145/3656019.3676900.
@article{jang2024rethinking,
title = {{Rethinking Page Table Structure for Fast Address Translation in GPUs: A Fixed-Size Hashed Page Table}},
author = {Jang, Sungbin and Park, Junhyeok and Kwon, Osang and Lee, Yongho and Hong, Seokin},
pages = {325--337},
year = {2024},
doi = {10.1145/3656019.3676900}
}Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.