TY - GEN
T1 - XaaS Containers
T2 - 2025 International Conference for High Performance Computing, Networking, Storage, and Analysis, SC 2025
AU - Copik, Marcin
AU - Alnuaimi, Eiman
AU - Kamatar, Alok
AU - Hayot-Sasson, Valerie
AU - Madonna, Alberto
AU - Gamblin, Todd
AU - Chard, Kyle
AU - Foster, Ian
AU - Hoefler, Torsten
N1 - Publisher Copyright:
© 2025 Copyright held by the owner/author(s).
PY - 2025/11/15
Y1 - 2025/11/15
N2 - High-performance computing (HPC) systems and cloud data centers are converging, and containers are becoming the default method of portable software deployment. Yet, while containers simplify software management, they face significant performance challenges in HPC environments as they must sacrifice hardware-specific optimizations to achieve portability. Although HPC containers can use runtime hooks to access optimized MPI libraries and GPU devices, they are limited by application binary interface (ABI) compatibility and cannot overcome the effects of early-stage compilation decisions. Acceleration as a Service (XaaS) proposes a vision of performance-portable containers, where a containerized application should achieve peak performance across all HPC systems. We present a practical realization of this vision through Source and Intermediate Representation (IR) containers, where we delay performance-critical decisions until the target system specification is known. We analyze specialization mechanisms in HPC software and propose a new LLM-assisted method for automatic discovery of specializations. By examining the compilation pipeline, we develop a methodology to build containers optimized for target architectures at deployment time. Our prototype demonstrates that new XaaS containers combine the convenience of containerization with the performance benefits of system-specialized builds.
AB - High-performance computing (HPC) systems and cloud data centers are converging, and containers are becoming the default method of portable software deployment. Yet, while containers simplify software management, they face significant performance challenges in HPC environments as they must sacrifice hardware-specific optimizations to achieve portability. Although HPC containers can use runtime hooks to access optimized MPI libraries and GPU devices, they are limited by application binary interface (ABI) compatibility and cannot overcome the effects of early-stage compilation decisions. Acceleration as a Service (XaaS) proposes a vision of performance-portable containers, where a containerized application should achieve peak performance across all HPC systems. We present a practical realization of this vision through Source and Intermediate Representation (IR) containers, where we delay performance-critical decisions until the target system specification is known. We analyze specialization mechanisms in HPC software and propose a new LLM-assisted method for automatic discovery of specializations. By examining the compilation pipeline, we develop a methodology to build containers optimized for target architectures at deployment time. Our prototype demonstrates that new XaaS containers combine the convenience of containerization with the performance benefits of system-specialized builds.
KW - Containers
KW - Intermediate Representation
KW - Performance Portability
UR - https://www.scopus.com/pages/publications/105023963994
U2 - 10.1145/3712285.3759868
DO - 10.1145/3712285.3759868
M3 - Contribution to conference proceedings
AN - SCOPUS:105023963994
T3 - Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, SC 2025
SP - 533
EP - 555
BT - Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, SC 2025
PB - Association for Computing Machinery, Inc
Y2 - 16 November 2025 through 21 November 2025
ER -