General Information
Details
Santosh Pande's research involves developing new compiler analyses and optimizations geared towards systems ranging from high performance (cloud computing) to specialized ML accelerators to energy efficient embedded processors.His research targets multiple dimensions of modern systems' performance such as speed, energy consumption, code size, security and real time aspects.
in past, he has done extensive work on energy efficient and compact code generation (register allocation in particular) for embedded processors such as ARM, network, DSP and reconfigurable processors resulting in publications in LCTES, PLDI and journals such as TECS, TPDS and TOPLAS.
More recently, his work on high performance (cloud) computing has focussed on how to assist schedulers to most efficiently share critical resources such as GPUs (refer to PPoPP 2022 and GPGPU 2022 papers) and last level caches (refer to PACT 2022 and OOPSLA 2023). In particular, in these works, he proposed a new concept of Beacons that generate dynamic loop characteristic at the entrances of loop nests to equip scheduler in terms of avoiding resource conflicts. The net result of such techniues is 2-4X improvement in the resource utilizations and throughput.
In terms of enhancing the security properties of large scale applications, he has developed a series of techniques for attack surface reduction by devloping purely static as well as predictive dynamic techniques to determine and just in time enable execution frontier of functions by using OS mechanisms. The net result being over 95% reduction on gadgets and elimination of all known CVEs in libraries such as glibc and servers such as nginx (refer to PLDI 2000 as well as ASPLOS 2025 papers).
His recent interests are how to use and optimize LLMs towards improving compiler optiizations.
Santosh Pande's teaching interests involve compilers and embedded software. His teaching philosophy involves building rigorous foundation and enhance it through combination of theory coupled with practice. The hope is to build curiosity and deep interest in the field leading to a pursuit of advanced practice or higher education as the case may be. He teaches compiler theory and practice and embedded software optimizations classes in the OMSCS (Online MS in CS) program and advanced compiler optimizations class on campus. He recruits highly motivated students from respective classes for pursuing state of the art research on frontier research topics.
Select Recent Publications: (full list of publications found at : https://dblp.org/pid/08/6824.html)
[Security] Chris Porter, Sharjeel Khan, Kangqi Ni, and Santosh Pande. 2025. Tackling ML-based Dynamic Mispredictions using Statically Computed Invariants for Attack Surface Reduction. In Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 (ASPLOS '25). Association for Computing Machinery, New York, NY, USA, 1218–1234. https://doi.org/10.1145/3676641.3716276
[Resource Sharing in Cloud and HPC Clusters] Girish Mururu, Sharjeel Khan, Bodhisatwa Chatterjee, Chao Chen, Chris Porter, Ada Gavrilovska, and Santosh Pande. 2023. Beacons: An End-to-End Compiler Framework for Predicting and Utilizing Dynamic Loop Characteristics. Proc. ACM Program. Lang. 7, OOPSLA2, Article 228 (October 2023), 31 pages. https://doi.org/10.1145/3622803
[Security] Sharjeel Khan, Bodhisatwa Chatterjee, and Santosh Pande. 2024. Pythia: Compiler-Guided Defense Against Non-Control Data Attacks. In Proceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3 (ASPLOS '24), Vol. 3. Association for Computing Machinery, New York, NY, USA, 850–866. https://doi.org/10.1145/3620666.3651343
[GPU sharing] Chao Chen, Chris Porter, and Santosh Pande. 2022. CASE: a compiler-assisted SchEduling framework for multi-GPU systems. In Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP '22). Association for Computing Machinery, New York, NY, USA, 17–31. https://doi.org/10.1145/3503221.3508423
[Cache Sharing] Bodhisatwa Chatterjee, Sharjeel Khan, and Santosh Pande. 2023. Com-CAS: Effective Cache Apportioning under Compiler Guidance. In Proceedings of the International Conference on Parallel Architectures and Compilation Techniques (PACT '22). Association for Computing Machinery, New York, NY, USA, 14–27. https://doi.org/10.1145/3559009.3569645
[Deman Drievn Verification] Sharjeel Khan, Bodhisatwa Chatterjee, and Santosh Pande. 2022. VICO: demand-driven verification for improving compiler optimizations. In Proceedings of the 36th ACM International Conference on Supercomputing (ICS '22). Association for Computing Machinery, New York, NY, USA, Article 16, 1–14. https://doi.org/10.1145/3524059.3532393
[Security] Chris Porter, Girish Mururu, Prithayan Barua, and Santosh Pande. 2020. BlankIt library debloating: getting what you want instead of cutting what you don’t. In Proceedings of the 41st ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI 2020). Association for Computing Machinery, New York, NY, USA, 164–180. https://doi.org/10.1145/3385412.3386017
[Soft real time systems] T. Kumar, K. Ni and S. Pande, "Characterizing Dominant Program Behavior Using the Execution-Time Variance of the Call Structure," 2019 IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS), Montreal, QC, Canada, 2019, pp. 117-129, doi: 10.1109/RTAS.2019.00018.