Managing Scalable Direct Storage Accesses for GPUs with GoFS
Shaobo Li, Yirui Eric Zhou, Yuqi Xue, Yuan Xu, Jian Huang
2025Year
2Top-tier citations
Abstract
As we shift from CPU-centric computing to GPU-accelerated computing for supporting intelligent data processing at scale, the storage bottleneck has been exacerbated. To bypass the host CPUand alleviate unnecessary data movements, modern GPUs enable direct storage access to SSDs (i.e., GPUDirect Storage). However, current GPUDirect Storage solutions still rely on the host file system to manage the storage device, direct storage accesses are still bottlenecked by the host.
Ask about this paper
Ask your agent about it.
Lune has read the top-tier papers around this one, so every answer names the papers it rests on.
Your agent calls
Lunesearch_papers
Free to start. No credit card required.
Terminal
Install the CLIlune papers get a4e5aa89-e256-4536-9dba-34fa72944a3cCited by top-tier papers2
- Cost-Efficient LLM Training with Lifetime-Aware Tensor Offloading via GPUDirect StorageZiqi Yuan, Haoyang Zhang, Yirui Eric Zhou, Apoorve Mohan et al.NeurIPS 2025 · 7 citations
- CoPilotIO: CPU as a Co-Pilot for GPU I/O to Free GPU ComputeGuanyi Chen, Qi Chen, Shu Yin, Jian ZhangOSDI 2026
Related papers
- CAM: Asynchronous GPU-Initiated, CPU-Managed SSD Management for Batching Storage AccessZiyu Song, Jie Zhang, Jie Sun, Mo Sun et al.ICDE 2025 · 4 citations
- GeminiFS: A Companion File System for GPUsShi Qiu, Weinan Liu, Yifan Hu, Jianqin Yan et al.FAST 2025 · 17 citations
- Phoenix: A Refactored I/O Stack for GPU Direct Storage without Phony BuffersJianqin Yan, Shi Qiu, Yina Lv, Yifan Hu et al.SC 2025 · 3 citations
- OS2G: A High-Performance DPU Offloading Architecture for GPU-based Deep Learning with Object StorageZhen Jin, Yiquan Chen, Mingxu Liang, Yijing Wang et al.ASPLOS 2025 · 5 citations
- Asynchrony and GPUs: Bridging this Dichotomy for I/O with AGIOJihoon Han, Anand Sivasubramaniam, Chia-Hao Chang, Vikram Sharma Mailthody et al.ASPLOS 2026 · 1 citation
