Data Movement-Aware GPU Sharing for Data-Intensive Systems

Summary: Introduces GUST, a GPU scheduler that treats PCIe as a first-class schedulable resource and classifies kernels as transfer- or device-intensive to interleave analytics (PCIe-heavy) and inference (compute-heavy) workloads. Prototype colocating four mixed workloads reduces performance degradation vs dedicated GPUs from 3.9–7x to 2.8x. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: 572
Venue: CIDR
Year: 2026
Pagerank: 5.093636e-05
Overall Rank: 10,119 | 30.58%
DOI: -

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Yi Jiang (EPFL)
2. Hamish Nicholson (EPFL)
3. Viktor Sanca (EPFL; Oracle)
4. Anastasia Ailamaki (EPFL)

BibTeX Citation

@inproceedings{jiang_cidr26,
        address = {Amsterdam, Netherlands},
        series = {{CIDR} '26},
        title = {{Data Movement-Aware GPU Sharing for Data-Intensive Systems}},
        booktitle = {Proceedings of the {Conference} on {Innovative} {Data} {Systems} {Research}},
        author = {Jiang, Yi and Nicholson, Hamish and Sanca, Viktor and Ailamaki, Anastasia},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 10 of 10 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
1,245	Palimpzest: Optimizing AI-Powered Analytics with Declarative Query Processing	2025	CIDR	0.00011507415
1,466	A Study of the Fundamental Performance Characteristics of GPUs and CPUs for Database Analytics	2020	SIGMOD	0.0001068941
1,977	HetExchange: Encapsulating heterogeneous CPU-GPU parallelism in JIT compiled engines	2019	VLDB	9.3641101e-05
2,809	Text2SQL is Not Enough: Unifying AI and Databases with TAG	2025	CIDR	8.0994951e-05
3,874	Tensors: An abstraction for general data processing	2021	VLDB	7.0561161e-05
5,308	Towards a Hybrid Design for Fast Query Processing in DB2 with BLU Acceleration Using Graphical Processing Units: A Technology Demonstration	2016	SIGMOD	6.2720812e-05
6,065	Vortex: Overcoming Memory Capacity Limitations in GPU-Accelerated Large-Scale Data Analytics	2025	VLDB	5.9895288e-05
6,835	GPU-accelerated data management under the test of time	2020	CIDR	5.7584924e-05
7,432	Powerful GPUs or Fast Interconnects: Analyzing Relational Workloads on Modern GPUs	2025	VLDB	5.6199783e-05
7,959	Terabyte-Scale Analytics in the Blink of an Eye	2026	VLDB	5.5181056e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	1,234	Large Graph Convolutional Network Training with GPU-Oriented Data Communication Architecture	2021	VLDB
2	6,774	Evaluating Multi-GPU Sorting with Modern Interconnects	2022	SIGMOD
3	10,985	Scaling GPU-Accelerated Databases beyond GPU Memory Size	2025	VLDB
4	7,432	Powerful GPUs or Fast Interconnects: Analyzing Relational Workloads on Modern GPUs	2025	VLDB
5	6,065	Vortex: Overcoming Memory Capacity Limitations in GPU-Accelerated Large-Scale Data Analytics	2025	VLDB
6	2,566	Pump Up the Volume: Processing Large Data on GPUs with Fast Interconnects	2020	SIGMOD
7	9,873	Workload Placement on Heterogeneous CPU-GPU Systems	2024	VLDB
8	4,607	Adaptive NUMA-aware data placement and task scheduling for analytical workloads in main-memory column-stores	2017	VLDB
9	2,504	Concurrent Analytical Query Processing with GPUs	2014	VLDB
10	6,835	GPU-accelerated data management under the test of time	2020	CIDR