Pith. sign in

REVIEW 3 major objections 5 minor 34 references

GORIO: GPU-Centered Remote I/O for Graph ANNS over NVMe-oF

T0 review · 3 major / 5 minor · reviewed 2026-07-11 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read GORIO keeps graph ANNS search ownership on the GPU over remote NVMe-oF, with the CPU only as a transport proxy, and reports 1.31× and 4.89× speedups on SIFT1M.

desk verdict Solid systems report with a coherent GPU-owned remote miss path; the 1.31×/4.89× numbers are real on SIFT1M but do not yet prove the large-index disaggregation story the intro sells. read the letter →

arxiv 2607.04415 v2 pith:62BQDQPA submitted 2026-07-05 cs.DC cs.DB

Gen Zhang , Wenhao Gu , Shan Huang , Xinhai Chen This is my paper · ORCID
classification cs.DCcs.DB
keywords ANNSvectordatabasesGPU-centricstorageNVMe-oFGPU-Directdisaggregatedgraphtraversalpagecache
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Large graph indexes for approximate nearest-neighbor search often do not fit in one GPU’s memory, so production systems want storage disaggregated over NVMe-oF. Existing remote paths still let the CPU form I/O, drive the network, and decide when GPU work may resume—exactly the wrong place for control, because the next page miss is discovered inside GPU graph traversal. GORIO extends GPU-centered local page-cache I/O to remote storage: GPU code still owns query evolution, miss generation, pending state, and resume decisions, while the CPU is reduced to submitting NVMe-oF commands and publishing completion bits into GPU-visible memory. A second layer adds ANNS-specific persistent scheduling so blocked queries yield and other ready work continues while remote pages arrive. On a SIFT1M DiskANN-style workload over RDMA NVMe-oF, the system is 1.31× faster than a state-of-the-art remote reference path and 4.89× faster than a direct remote page-cache path that blocks on each miss. The claim is that this split of ownership is both workable and necessary for irregular GPU graph search over fabric storage.

What carries the argument

Split-phase remote page service: a GPU page-cache miss becomes a lock-free request descriptor, the blocked query yields, independent graph work continues, and a GPU-visible ready-table entry lets the persistent scheduler resume when the CPU proxy has filled the cache slot over NVMe-oF.

What would settle it

Re-run the same five paths on a multi-billion-vector DiskANN-style index with deeper traversal and a working set far larger than GPU memory; if GORIO’s 1.31× and 4.89× advantages disappear or reverse under that miss load, the substrate claim does not hold for the intended setting.

Watch

Extended reading notes

Core claim

GPU-centered local page-cache I/O can be extended across real NVMe-oF so that graph ANNS keeps query evolution, page-miss generation, pending state, and resume decisions on the GPU, with the CPU restricted to transport and completion proxying; on SIFT1M this substrate is 1.31× faster than the remote-I/O reference path and 4.89× faster than a direct remote page-cache path that waits in place for pages.

Load-bearing premise

The paper’s speedups rest on single-run SIFT1M search times over one small index and one RDMA NVMe-oF link, which may not reproduce the miss pressure and cache behavior of the large disaggregated indexes the design is meant for.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. GORIO is a systems paper that extends GPU-centered local page-cache I/O (in the BaM style) to remote NVMe-oF for graph-based ANNS. The design keeps query evolution, page-miss generation, pending state, and resume decisions on the GPU, while the CPU is restricted to an SPDK/NVMe-oF transport and completion proxy that publishes ready state into GPU-visible memory. Two layers are claimed: a GPU-direct remote path that turns page-cache misses into split-phase remote operations (descriptors, ready table, cache fill), and ANNS-specific persistent GPU scheduling that yields blocked queries and resumes them when pages arrive. On a SIFT1M DiskANN-style workload (10k queries, recall@10) over RDMA NVMe-oF, the paper reports G-Gust 1.31× faster than GustANN and G-ABaM 4.89× faster than a BaM-style remote path, with large gains over a GDS baseline (Table 2, §5.2).

Significance. If the architectural split and speedups hold under realistic remote-miss pressure, the paper would be a useful contribution to GPU-centric storage and disaggregated vector search: it cleanly separates GPU ownership of irregular graph-search state from host-managed NVMe-oF progress, and the ablation structure (G-Gust vs GustANN for the remote backend; G-ABaM vs BaM for ANNS scheduling) is a sensible way to isolate those layers. The work is concrete rather than purely conceptual—it is implemented with CUDA/SPDK over real RDMA NVMe-oF and reports end-to-end search times against named baselines. That said, significance for the motivated large-index, capacity-disaggregated setting is currently limited by evaluation scale; the paper itself notes the need for larger indexes and datasets (§7.2).

major comments (3)
  1. Table 2 / §5.2 and abstract/Result Takeaway: the central claim that GORIO demonstrates a GPU-centered remote I/O substrate for graph ANNS rests on SIFT1M (1M vectors, 10k queries, one L40S, one InfiniBand path, one enterprise SSD). §1–2 and Motivations 1–2 motivate the work by large indexes that exceed GPU memory and by fine-grained, data-dependent remote page misses. On a small index the GPU page cache can absorb much of the working set after warm-up, so the 1.31× and 4.89× gains may largely reflect local scheduling/proxy overhead rather than fabric-crossing miss handling under capacity pressure. §7.2 itself states that larger indexes and additional datasets are needed for higher graph depth and larger working sets, yet the abstract, contributions, and takeaway still present SIFT1M as demonstrating the remote substrate. Without miss-rate, cache-hit, outstanding-I/O depth, or working-set
  2. §5.1–5.2: all end-to-end times appear to be single-run application search times with no variance, repeated trials, or sensitivity to free parameters the design depends on (CPU proxy batching and outstanding depth, GPU page-cache size/slot policy, scheduler yield/claim policy). For a systems result whose headline numbers are 1.31× and 4.89×, single-point timings are weak evidence of a stable substrate advantage. Please report multi-run means/std or ranges, and at least a brief sensitivity study on cache size and outstanding remote I/O depth so readers can see whether the gains are robust or tuned to one configuration.
  3. §5.1 Comparison and Table 2: the paper asserts “same recall” in the conclusion and uses recall@10 as the evaluation setting, but Table 2 reports only time/QPS/speedup with no Recall@10 column or confirmation that all five paths achieve the same recall on the same index. If any path trades recall for latency (e.g., via different pending/resume or page-service behavior), the throughput comparison is not apples-to-apples. Please report Recall@10 for every system in Table 2 and state the common search parameters (ef/beam, graph degree, page size) used across paths.
minor comments (5)
  1. Figure 1 is described in §3.1 but the manuscript text does not include quantitative labels (e.g., where descriptors live, ready-bit protocol). A short caption note on the ready-table vs. request-queue handoff would help readers who only skim the figure.
  2. §4 Implementation: “GustANN-derived” and “BaM-based” are clear at a high level, but the text does not state which BaM/GustANN components were reused vs. rewritten (page cache, scheduler, SPDK path). A short reuse table would improve reproducibility claims.
  3. Table 1 lists CUDA 13.0 and SPDK 25.09; if these are intentional, fine, but double-check version strings for the camera-ready so readers can match the stack.
  4. Related work (§6) is appropriate but could more explicitly contrast GORIO with concurrent GPU-storage ANNS systems that appear in the reference list (e.g., FlashANNS, Neos) on the remote vs. local axis.
  5. Minor wording: abstract and §1 use both “GORIO” and spaced forms inconsistently in places; unify the product name and the G-Gust / G-ABaM labels early in §5.

Circularity Check

0 steps flagged · score 0.0 of 10

Empirical systems paper: speedups are measured against external baselines, not derived from fitted parameters or self-defined identities.

full rationale

GORIO’s load-bearing claims are end-to-end measurements (Table 2): G-Gust 0.99 s vs GustANN 1.30 s (1.31×) and G-ABaM 1.77 s vs BaM 8.66 s (4.89×) on SIFT1M over RDMA NVMe-oF. These are experimental comparisons against named reference paths (GustANN-style SPDK remote I/O, BaM-style demand paging, GDS), not predictions obtained by fitting a parameter to the same quantity or by defining X in terms of Y. The design narrative (GPU owns miss generation/pending/resume; CPU is transport proxy; split-phase remote page service) is an architectural specialization of prior local GPU I/O (BaM) and ANNS scheduling ideas; citing BaM/DiskANN/SPDK is normal prior-art framing and is not used as a uniqueness theorem or self-citation chain that forces the result. There are no equations that reduce a claimed prediction to its inputs by construction, no fitted-input-as-prediction steps, and no load-bearing self-citation of an unverified uniqueness result. Weaknesses of the evaluation (SIFT1M scale, single-run times) are external-validity concerns, not circularity. Score 0 is appropriate.

Assumptions & free parameters 3 free parameters · 4 assumptions · 3 invented entities

The central claims rest on standard systems assumptions plus one load-bearing experimental scope choice: that SIFT1M single-run timings on one RDMA NVMe-oF pair represent the intended disaggregated ANNS setting. No physical free parameters are fitted; invented entities are engineering modules, not new physical objects.

free parameters (3)
  • CPU proxy batching and outstanding remote I/O depth
    Described as controlled by the proxy (§3.3, §4) but no values, sweep, or sensitivity analysis are reported; throughput claims can depend on these knobs.
  • GPU page-cache size / slot allocation policy
    Miss rate and resume behavior depend on cache capacity and replacement; configuration is not quantified in the evaluation.
  • Persistent scheduler work-claim / yield policy
    Overlap gains in G-ABaM depend on how blocks claim ready queries and yield on misses; policy details are qualitative only.
assumptions (4)
  • domain assumption NVMe-oF/SPDK queue-pair progress must be host-managed, so a CPU transport proxy is required.
    Stated in §3.3 and §4 as the reason the CPU remains in the path; standard for current SPDK/NVMe-oF stacks.
  • domain assumption GPU-visible ready-table bits plus cache-slot fills are a correct completion interface for resuming pending graph queries without a shared GPU completion queue.
    Core of the remote I/O design in §3.2; correctness of resume depends on this visibility model.
  • ad hoc to paper SIFT1M DiskANN-style graph with 10k queries and recall@10 is a representative benchmark for the remote graph-ANNS substrate claims.
    All speedups in §5 use this single workload; §7.2 admits larger indexes are needed.
  • domain assumption BaM-style GPU page-cache miss generation is an appropriate base model to extend from local NVMe to remote NVMe-oF.
    Design premise throughout §2–§3; inherits BaM’s page-cache abstraction.
invented entities (3)
  • GORIO GPU-direct remote I/O path (miss → descriptor → SPDK NVMe-oF → ready table)
    purpose: Convert GPU page-cache misses into split-phase remote operations with GPU-visible completion.
    Primary system mechanism; engineering composition of known I/O primitives rather than a new physical entity.
  • ANNS-specific persistent GPU scheduler with pending/resume queues
    purpose: Overlap graph traversal with remote page service instead of blocking on misses.
    Explains the 4.89× G-ABaM vs BaM gain; scheduler is paper-specific specialization.
  • Thin CPU NVMe-oF transport/completion proxy outside graph-search semantics
    purpose: Advance host-managed transport without taking ownership of candidate queues or query termination.
    Defines the ownership split that the paper argues is necessary for GPU-centered remote ANNS.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GORIO: GPU-Centered Remote I/O for Graph ANNS over NVMe-oF." pith.science (2026). https://pith.science/paper/62BQDQPA

@misc{pith2026260704415,
  author       = {Pith},
  title        = {Pith review of: GORIO: GPU-Centered Remote I/O for Graph ANNS over NVMe-oF},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/62BQDQPA}},
  note         = {Machine review of arXiv:2607.04415}
}
read the original abstract

Graph-based approximate nearest neighbor search (ANNS) is increasingly used in vector databases and retrieval-augmented generation services, but large vector indexes often exceed the memory capacity of a single GPU server. NVMe over Fabrics (NVMe-oF) provides an attractive storage-disaggregation substrate, yet existing remote storage paths are still largely CPU-centered: the CPU forms I/O requests, drives transport progress, and determines when GPU computation can resume. This organization is poorly matched to graph ANNS, where the next data access is discovered inside GPU graph traversal. This paper presents GORIO, a system study that extends GPU-centered local I/O to remote storage and specializes the resulting substrate for graph ANNS over NVMe-oF. GORIO keeps query evolution, page-miss generation, pending-query state, and resume decisions on the GPU, while the CPU acts only as an NVMe-oF transport and completion proxy. The design has two layers: a GPU-direct remote I/O path that turns local page-cache misses into split-phase remote operations, and ANNS-specific scheduling mechanisms that overlap graph traversal with remote page service. On a SIFT1M DiskANN-style graph workload over an RDMA NVMe-oF path, GORIO is 1.31X faster than the state-of-the-art remote-I/O reference path and 4.89X faster than the direct remote page-cache path. These results demonstrate a concrete GPU-centered remote I/O substrate for graph ANNS.

Figures

Figures reproduced from arXiv: 2607.04415 by the authors.

Figure 1
Figure 1. GORIO architecture. The GPU owns graph-search state, page-miss generation, pending queues, and resume decisions. CPU workers are restricted to NVMe-oF request submission, polling, and completion publication into GPU-visible page-cache state. request allocation and dependency tracking on the GPU side in￾stead of serializing miss generation through a CPU-owned submit path. Second, CPU proxy workers consume ready descr… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 4 canonical work pages

  1. [1]

    Chia-Hao Chang, Jihoon Han, Anand Sivasubramaniam, Vikram Sharma Mailthody, Zaid Qureshi, and Wen-Mei Hwu. 2024. GMT: GPU Orchestrated Memory Tiering for the Big Data Era. InProceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3 (ASPLOS ’24). doi:10.1145/3620666.3651353

    doi:10.1145/3620666.3651353 2024

  2. [2]

    GNStor Authors. 2026. GNStor: Design of GPU-Native High-Performance Remote All-Flash Array. Local PDF copy; bibliographic metadata unavailable

    2026

  3. [3]

    Jihoon Han, Anand Sivasubramaniam, Chia-Hao Chang, Vikram Sharma Mailthody, Zaid Qureshi, and Wen-Mei Hwu. 2026. Asynchrony and GPUs: Bridging this Dichotomy for I/O with AGIO. InProceedings of the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 (ASPLOS ’26). doi:10.1145/3779212.3790130

    doi:10.1145/3779212.3790130 2026

  4. [4]

    Yifan Hu, Shi Qiu, Jianqin Yan, Hao Chen, Xintao Wang, Lu Tang, Guangtao Xue, and Yiming Zhang. 2025. TARDIS: A GPU-Centric KV Cache Service for Efficient LLM Inference. InProceedings of the 16th ACM SIGOPS Asia-Pacific Workshop on Systems (APSys ’25)

    2025

  5. [5]

    Yuchen Huang, Xiaopeng Fan, Song Yan, and Chuliang Weng. 2024. Neos: A NVMe-GPUs Direct Vector Service Buffer in User Space. InProceedings of the 40th IEEE International Conference on Data Engineering (ICDE ’24). doi:10.1109/ ICDE60146.2024.00289

    arXiv 2024

  6. [6]

    Haodi Jiang, Hao Guo, Minhui Xie, Jiwu Shu, and Youyou Lu. 2025. High- Throughput, Cost-Effective Billion-Scale Vector Search with a Single GPU.Pro- ceedings of the ACM on Management of Data3, 6, Article 334 (2025). doi:10.1145/ 3769799

    2025

  7. [7]

    2022.Creating Large Real and Synthetic Graph Datasets for GNN Applications

    Arpandeep Khatua. 2022.Creating Large Real and Synthetic Graph Datasets for GNN Applications. Master’s thesis. University of Illinois Urbana-Champaign

    2022

  8. [8]

    Shaobo Li, Yirui Eric Zhou, Yuqi Xue, Yuan Xu, and Jian Huang. 2025. Managing Scalable Direct Storage Accesses for GPUs with GoFS. InProceedings of the 31st ACM Symposium on Operating Systems Principles (SOSP ’25)

    2025

Show all 34 references
  1. [9]

    Zhonggen Li, Xiangyu Ke, Yifan Zhu, Yunjun Gao, and Feifei Li. 2026. Efficient Graph Embedding at Scale: Optimizing CPU-GPU-SSD Integration.arXiv preprint arXiv:2505.09258(2026). https://arxiv.org/abs/2505.09258

    2026

  2. [10]

    Xiaojian Liao, Youyou Lu, Zhe Yang, and Jiwu Shu. 2023. Efficient Crash Con- sistency for NVMe over PCIe and RDMA.ACM Transactions on Storage19, 1, Article 7 (2023). doi:10.1145/3568428

    2023 doi

  3. [11]

    2022.Application Support and Adaptation for High- Throughput Accelerator Orchestrated Fine-Grain Storage Access

    Vikram Sharma Mailthody. 2022.Application Support and Adaptation for High- Throughput Accelerator Orchestrated Fine-Grain Storage Access. Ph. D. Dissertation. University of Illinois Urbana-Champaign

    2022

  4. [12]

    Malkov and Dmitry A

    Yury A. Malkov and Dmitry A. Yashunin. 2020. Efficient and Robust Approximate Nearest Neighbor Search Using Hierarchical Navigable Small World Graphs.IEEE Transactions on Pattern Analysis and Machine Intelligence42, 4 (2020), 824–836. doi:10.1109/TPAMI.2018.2889473

    2020 doi

  5. [13]

    Seung Won Min, Kun Wu, Sitao Huang, Mert Hidayetoglu, Jinjun Xiong, Eiman Ebrahimi, Deming Chen, and Wen mei Hwu. 2021. PyTorch-Direct: Enabling GPU Centric Data Access for Very Large Graph Neural Network Training with Irregular Accesses.arXiv preprint arXiv:2101.07956(2021). ...

    2021 arXiv

  6. [14]

    NVM Express. 2021. NVM Express over Fabrics Specification, Revision 1.1a. https://nvmexpress.org/wp-content/uploads/NVMe-over-Fabrics-1.1a- 2021.07.12-Ratified.pdf. Ratified July 12, 2021; accessed for NVMe-oF trans- port model and terminology

    2021

  7. [15]

    Jeongmin Brian Park, Vikram Sharma Mailthody, Zaid Qureshi, and Wen mei Hwu. 2024. GIDS: Accelerating Sampling and Aggregation Operations in GNN Frameworks with GPU Initiated Direct Storage Accesses.Proceedings of the VLDB Endowment17, 6 (2024), 1227–1240. doi:10.14778/3648160.3648166

    2024 doi

  8. [16]

    Phoenix Authors. 2025. Phoenix: A Refactored I/O Stack for GPU Direct Storage without Phony Buffers. InProceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis (SC ’25). doi:10.1145/ 3689031.3696080 Author names are not avai...

    2025

  9. [17]

    Shi Qiu, Weinan Liu, Yifan Hu, Jianqin Yan, Zhirong Shen, Xin Yao, Renhai Chen, Gong Zhang, and Yiming Zhang. 2025. GeminiFS: A Companion File System for GPUs. InProceedings of the 23rd USENIX Conference on File and Storage Technologies (FAST ’25)

    2025

  10. [18]

    2022.Infrastructure to Enable and Exploit GPU Orchestrated High- Throughput Storage Access

    Zaid Qureshi. 2022.Infrastructure to Enable and Exploit GPU Orchestrated High- Throughput Storage Access. Ph. D. Dissertation. University of Illinois Urbana- Champaign

    2022

  11. [19]

    Zaid Qureshi, Vikram Sharma Mailthody, Isaac Gelado, Seungwon Min, Amna Masood, Jeongmin Park, Jinjun Xiong, C. J. Newburn, Dmitri Vainbrand, I.-Hsin Chung, Michael Garland, William Dally, and Wen mei Hwu. 2023. GPU-Initiated On-Demand High-Throughput Storage Access in the BaM...

    2023 doi

  12. [20]

    Ziyu Song, Jie Zhang, Jie Sun, Mo Sun, Zihan Yang, Zheng Zhang, Xuzheng Chen, Fei Wu, Huajin Tang, and Zeke Wang. 2025. CAM: Asynchronous GPU-Initiated, CPU-Managed SSD Management for Batching Storage Access. InProceedings of the 41st IEEE International Conference on Data Engi...

    2025

  13. [21]

    SPDK Project. 2026. Storage Performance Development Kit (SPDK). https: //spdk.io/. Accessed for system software description

    2026

  14. [22]

    Suhas Jayaram Subramanya, Rohan Kadekodi, Ravishankar Krishnaswamy, and Harsha Vardhan Simhadri. 2019. DiskANN: Fast Accurate Billion-Point Nearest Neighbor Search on a Single Node. InAdvances in Neural Information Processing 5 Technical Report, 2026, Zhang et al. Systems 32 (...

    2019

  15. [23]

    Jie Sun, Mo Sun, Zheng Zhang, Zuocheng Shi, Jun Xie, Zihan Yang, Jie Zhang, Zeke Wang, and Fei Wu. 2025. Hyperion: Co-Optimizing SSD Access and GPU Computation for Cost-Efficient GNN Training. InProceedings of the 41st IEEE International Conference on Data Engineering (ICDE ’2...

    2025 doi

  16. [24]

    Jie Sun, Mo Sun, Zheng Zhang, Jun Xie, Zuocheng Shi, Zihan Yang, Jie Zhang, Fei Wu, and Zeke Wang. 2023. Helios: An Efficient Out-of-core GNN Training System on Terabyte-scale Graphs with In-memory Performance.arXiv preprint arXiv:2310.00837(2023). https://arxiv.org/abs/2310.00837

    2023 arXiv

  17. [25]

    Bing Tian, Haikun Liu, Yuhang Tang, Shihai Xiao, Zhuohui Duan, Xiaofei Liao, Hai Jin, Xuecang Zhang, Junhua Zhu, and Yu Zhang. 2025. Towards High- throughput and Low-latency Billion-scale Vector Search via CPU/GPU Collabo- rative Filtering and Re-ranking. InProceedings of the ...

    2025

  18. [26]

    Karthik Venkatasubba, Saim Khan, Somesh Singh, Harsha Vardhan Simhadri, and Jyothi Vedurada. 2025. BANG: Billion-Scale Approximate Nearest Neighbour Search Using a Single GPU.IEEE Transactions on Big Data11, 6 (2025), 3142–3157. doi:10.1109/TBDATA.2025.3581085

    2025 doi

  19. [27]

    Mengzhao Wang, Weizhi Xu, Xiaomeng Yi, Songlin Wu, Zhangyang Peng, Xi- angyu Ke, Yunjun Gao, Xiaoliang Xu, Rentong Guo, and Charles Xie. 2024. Starling: An I/O-Efficient Disk-Resident Graph Index Framework for High- Dimensional Vector Similarity Search on Data Segment.Proceedi...

    2024 doi

  20. [28]

    Yang Xiao, Mo Sun, Ziyu Song, Bing Tian, Jie Sun, Jie Zhang, Zeke Wang, Zonghui Wang, Wenzhi Chen, and Fei Wu. 2026. FlashANNS: GPU-Driven Asynchronous I/O Pipelining for Eliminating Storage-Compute Bottlenecks in Billion-Scale Similarity Search.Proceedings of the ACM on Manag...

    2026

  21. [29]

    Minhui Xie, Youyou Lu, Yangyang Feng, and Jiwu Shu. 2024. A Recommendation Model Inference System Based on GPU Direct Storage Access Architecture. Journal of Computer Research and Development61, 3 (2024), 589–599. doi:10.7544/ issn1000-1239.202330402

    2024

  22. [30]

    Jones, and Peipei Zhou

    Zhuoping Yang, Jinming Zhuang, Xingzhen Chen, Alex K. Jones, and Peipei Zhou

  23. [31]

    InProceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC ’25)

    AGILE: Lightweight and Efficient Asynchronous GPU-SSD Integration. InProceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC ’25). doi:10.1145/3712285.3759778

    doi

  24. [32]

    Zili Zhang, Fangyue Liu, Gang Huang, Xuanzhe Liu, and Xin Jin. 2024. Fast Vector Query Processing for Large Datasets Beyond GPU Memory with Reordered Pipelining. InProceedings of the 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI ’24)

    2024

  25. [33]

    Hao Zhou, Yuanhui Chen, Wu Zeng, Lixiao Cui, Gang Wang, and Xiaoguang Liu. 2025. GPComp: Using GPU and SSD-GPU Peer to Peer DMA to Accelerate LSM-Tree Compaction for Key-Value Store.IEEE Transactions on Parallel and Distributed Systems36, 9 (2025), 1920–1936. doi:10.1109/TPDS....

    2025 doi

  26. [34]

    2025.Managing Scalable Direct Storage Accesses for GPUs

    Yirui Eric Zhou. 2025.Managing Scalable Direct Storage Accesses for GPUs. Mas- ter’s thesis. University of Illinois Urbana-Champaign. 6

    2025

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.