A Data-Driven Dynamic Execution Orchestration Architecture Conference: Proceedings of the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1; March 22–26, 2026; Pittsburgh, PA, USA Proceedings of the 31st ACM International Conference

Zhenyu Bai, Pranav Dangi, Rohan Juneja, Zhaoying Li, Zhanglu Yan, Huiying Lan, Tulika Mitra · 2026

Read A Data-Driven Dynamic Execution Orchestration Architecture Conference: Proceedings of the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1; March 22–26, 2026; Pittsburgh, PA, USA Proceedings of the 31st ACM International Conference by Zhenyu Bai, Pranav Dangi, Rohan Juneja, Zhaoying Li, Zhanglu Yan, Huiying Lan, Tulika Mitra (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-17119.txt.

Description

Zhenyu Bai led a seven-author team designing a processor architecture that refuses to choose between specialization and flexibility. Domain-specific accelerators excel on their target workloads but falter on new kernels or ragged inputs, while programmable chips like FPGAs and GPUs stay flexible yet lose efficiency on irregular, sparse data. The paper introduces Canon, a parallel architecture meant to match specialist efficiency without giving up general-purpose agility.

Canon runs on dynamic, data-driven control: programmable finite state machines, set at compile time, turn incoming metadata such as sparse coordinates into control instructions on the fly. A time-lapsed SIMD scheme staggers instructions across processing units over several cycles, spreading control cost thin. Tests across varied kernels showed high performance with specialist-like efficiency, keeping the flexibility that programmable hardware promises.

Catalog record

CREATORZhenyu Bai, Pranav Dangi, Rohan Juneja, Zhaoying Li, Zhanglu Yan, Huiying Lan, Tulika Mitra
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE70 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-17119

Browse the full library · About the Tower of Babel