--- title: "Research and publications | IOWarp" description: "IOWarp research on HPC I/O, context infrastructure, and scientific AI agents, with a browsable index of papers and posters since 2022." url: https://iowarp.ai/research/ --- > IOWarp research on HPC I/O, context infrastructure, and scientific AI agents, with a browsable index of papers and posters since 2022. # Systems science for scientific AI. For researchers and reviewers: the systems research behind IOWarp, from memory and HPC I/O to context and scientific AI agents. The record includes papers from the IOWarp team, related research, and earlier Gnosis Research Center work the project builds on. Record Papers and posters, 2022–2026 Index Browse the publication index Feed [RSS feed of publications](https://iowarp.ai/research/rss.xml) Open questions [Research directions](https://iowarp.ai/research/directions/) At SC26 [Workshop presentations, research posters, the conference paper, and the HDF5 discussion](https://iowarp.ai/community/#sc26) Broader work [Gnosis Research Center publications ↗](https://grc.iit.edu/publications) The research question ## How does scientific data become useful context? More throughput alone is not enough. Workflows need data placement, provenance, access, and reasoning to survive changes in format, scale, and computational environment. The CLIO platform grows from that systems question. 01 / Through lines ## Performance, context, and agency. - 01 / Performance ### Move less. Move well. Memory and storage blending, tiered I/O, and dataflow-aware execution help scientific workloads use the resources they have. - 02 / Context ### Keep the meaning. Data becomes more useful when metadata, structure, and computational state remain available to the next step. - 03 / Agency ### Make tools usable. Scientific agents need grounded access to datasets, clusters, instruments, and evidence they can inspect. 02 / Selected papers ## A few starting points. - 01 / 2025 ### WisIO: Automated I/O Bottleneck Detection with Multi-Perspective Views for HPC Workflows ICS'25: 2025 International Conference on Supercomputing Read the paper ↗ - 02 / 2024 ### MegaMmap: Blurring the Boundary Between Memory and Storage for Data-Intensive Workloads Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis SC Read the paper ↗ - 03 / 2024 ### DFTracer: An Analysis-Friendly Data Flow Tracer for AI-Driven Workflows Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis Read the paper ↗ 03 / Publication index ## The archive. 1. 01 Findings of the Association for Computational Linguistics: EACL 2026 ### [PagedEviction: Structured Block-wise KV Cache Pruning for Efficient Large Language Model Inference](https://doi.org/10.18653/v1/2026.findings-eacl.168) K. T. Chitty-Venkata, J. Ye, S. Raskar, A. Kougkas, X.-H. Sun, M. Emani, V. Vishwanath, B. Nicolae 2. 02 Proceedings of the 35th International Symposium on High-Performance Parallel and Distributed Computing (HPDC), 2026 ### [PKAS: Predictive KVCache-Aware Scheduling for Faster LLM and Transformer Inferences](https://doi.org/10.1145/3806645.3807819) J. Ye, A. Maurya, K. T. Chitty-Venkata, B. Nicolae, A. Kougkas, X.-H. Sun 3. 03 2026 IEEE International Parallel and Distributed Processing Symposium (IPDPS) ### [Characterizing Dataflow for I/O-Aware Scheduling in HPC Workflows](https://doi.org/10.1109/IPDPS65963.2026.00076) M. Tang, L. Guo, A. Kougkas, X.-H. Sun, N. R. Tallent 4. 04 2026 IEEE International Conference on eScience (eScience) ### [AgentStage: Exploiting LLM Thinking for Data Staging in Scientific Agents](https://www.escience-conference.org/2026/papers) I. Yildirim, X.-H. Sun, A. Kougkas 5. 05 2026 IEEE International Conference on eScience (eScience) ### [Acropolis: A Portable, AI-Driven Semantic Search System for HPC](https://www.escience-conference.org/2026/papers) R. P. Pawar, L. Logan, J. Cernuda, X.-H. Sun, A. Kougkas 6. 06 International Conference for High Performance Computing, Networking, Storage, and Analysis (SC 2026) ### [Trigger, Render, Reason: Multimodal LLM Agents for Science-Aware Scientific Simulation Monitoring](https://grc.iit.edu/publications/xu-2026-trigger-render-reason-multimodal-llm-agents-c360/) H. Xu, J. Cernuda, A. Gainaru, S. Klasky, A. Kougkas, X.-H. Sun 7. 07 Conference on Language Modeling (COLM 2026) ### [InfiniteScienceGym: An Unbounded, Procedurally-Generated Benchmark for Scientific Analysis](https://arxiv.org/abs/2604.13201) O. Bentham, V. Srikumar 8. 08 ICS'25: 2025 International Conference on Supercomputing ### [WisIO: Automated I/O Bottleneck Detection with Multi-Perspective Views for HPC Workflows](https://doi.org/10.1145/3721145.3725742) I. Yildirim, H. Devarajan, A. Kougkas, X.-H. Sun, K. Mohror 9. 09 2025 IEEE International Conference on eScience (eScience) ### [Towards an AI-driven scientific workflow](https://doi.org/10.1109/eScience65000.2025.00074) S. Islam, J. Cernuda, I. Muradli, A. Kougkas, X.-H. Sun Poster 10. 10 2025 IEEE International Conference on eScience (eScience) ### [Insights into GPUDirect Data Transfer through NIXL Benchmarking](https://doi.org/10.1109/eScience65000.2025.00075) I. Muradli, J. Ye, L. Logan, A. Kougkas, X.-H. Sun Poster 11. 11 The International Parallel Data Systems Workshop (PDSW'24), 2024 ### Jarvis: Towards a Shared, User-Friendly, and Reproducible I/O Infrastructure J. Cernuda, L. Logan, N. Lewis, S. Byna, X.-H. Sun, A. Kougkas 12. 12 The International Conference for High Performance Computing, Networking, Storage, and Analysis (SC'24), 2024 ### Uncover the Overhead and Resource Usage for Handling KV Cache Overflow in LLM Inference J. Ye, B. Nicolae, A. Kougkas, X.-H. Sun Poster 13. 13 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis SC, 2024 ### [MegaMmap: Blurring the Boundary Between Memory and Storage for Data-Intensive Workloads](https://doi.org/10.1109/SC41406.2024.00114) L. Logan, A. Kougkas, X.-H. Sun 14. 14 Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, 2024 ### [DFTracer: An Analysis-Friendly Data Flow Tracer for AI-Driven Workflows](https://dl.acm.org/doi/abs/10.1109/SC41406.2024.00023) H. Devarajan, L. Pottier, K. Velusamy, H. Zheng, I. Yildirim, O. Kogiou, W. Yu, A. Kougkas, X.-H. Sun, J. S. Yeom, K. Mohror 15. 15 2024 IEEE 36th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD) ### [To Derive or Not to Derive: I/O Libraries Take Charge of Derived Quantities Computation](https://ieeexplore.ieee.org/document/10763877) A. Gainaru, N. Podhorszki, L. Dulac, Q. Gong, S. Klasky, G. Eisenhauer, A. Kougkas, X.-H. Sun, J. Lofstead 16. 16 Proceedings of the IEEE International Conference on Cluster Computing, 2024 ### [DaYu: Optimizing Distributed Scientific Workflows by Decoding Dataflow Semantics and Dynamics](https://ieeexplore.ieee.org/document/10740817) M. Tang, J. Cernuda, J. Ye, L. Guo, N. R. Tallent, A. Kougkas, X.-H. Sun 17. 17 Proceedings of the ACM SIGOPS Operating Systems Review, 2024 ### [An Evaluation of DAOS for Simulation and Deep Learning HPCWorkloads](https://doi.org/10.1145/3689051.3689058) L. Logan, J. Lofstead, X.-H. Sun, A. Kougkas 18. 18 Proceedings of the 53rd International Conference on Parallel Processing, 2024 ### [Viper: A High-Performance I/O Framework for Transparently Updating, Storing, and Transferring Deep Neural Network Models](https://doi.org/10.1145/3673038.3673070) J. Ye, J. Cernuda, N. Rajesh, K. Bateman, O. Yildiz, T. Peterka, A. Nigmetov, D. Morozov, X.-H. Sun, A. Kougkas, B. Nicolae 19. 19 Proceedings of the 53rd International Conference on Parallel Processing, 2024 ### [HStream: A hierarchical data streaming engine for high-throughput scientific applications](https://dl.acm.org/doi/abs/10.1145/3673038.3673150) J. Cernuda, J. Ye, A. Kougkas, X.-H. Sun 20. 20 Proceedings of the International Parallel and Distributed Processing Symposium, 2024 ### [TunIO: An AI-powered Framework for Optimizing HPC I/O](https://ieeexplore.ieee.org/document/10579249) N. Rajesh, K. Bateman, J. L. Bez, S. Byna, A. Kougkas, X.-H. Sun 21. 21 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2023 ### [Data Flow Lifecycles for Optimizing Workflow Coordination](https://dl.acm.org/doi/abs/10.1145/3581784.3607104) H. Lee, L. Guo, M. Tang, J. Firoz, N. Tallent, A. Kougkas, X.-H. Sun 22. 22 Proceedings of the SC'23 Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis, 2023 ### [IOMax: Maximizing Out-of-Core I/O Analysis Performance on HPC Systems](https://dl.acm.org/doi/abs/10.1145/3624062.3624191) I. Yildirim, H. Devarajan, A. Kougkas, X.-H. Sun, K. Mohror 23. 23 Proceedings of the 3rd Workshop on Challenges and Opportunities of Efficient and Performant Storage Systems, 2023 ### [An Evaluation of DAOS for Simulation and Deep Learning HPC Workloads](https://dl.acm.org/doi/abs/10.1145/3578353.3589542) L. Logan, J. Lofstead, X.-H. Sun, A. Kougkas 24. 24 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2022 ### [LabStor: A modular and extensible platform for developing high-performance, customized I/O stacks in userspace](https://ieeexplore.ieee.org/abstract/document/10046077) L. Logan, J. Cernuda, J. Lofstead, X.-H. Sun, A. Kougkas 25. 25 Proceedings of the 22nd IEEE International Symposium on Cluster, Cloud and Internet Computing, 2022 ### [Stimulus: Accelerate Data Management for Scientific AI applications in HPC](https://ieeexplore.ieee.org/abstract/document/9826104) H. Devarajan, A. Kougkas, H. Zheng, V. Vishwanath, X.-H. Sun Next step ## Follow the open questions. [Research directions](https://iowarp.ai/research/directions/) [See the demonstrations](https://iowarp.ai/research/demos/)