FHECore: Rethinking GPU Microarchitecture for Fully Homomorphic Encryption
2026
Read FHECore: Rethinking GPU Microarchitecture for Fully Homomorphic Encryption (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-2602-22229.txt.
Description
Fully homomorphic encryption lets computers work on encrypted data, but at enormous cost, often orders of magnitude slower than plain computation. Lohit Daksha and twelve co-authors argue GPUs, not custom chips, are the pragmatic platform: they are scalable, programmable, and widely available. The snag is that modern GPUs optimize for low-precision machine-learning math, a poor match for the wide-precision modular arithmetic encryption needs.
The paper proposes rethinking GPU design around that gap, showing how encryption primitives can exploit GPU parallelism despite it. For privacy-preserving computing, it sketches a path to encrypted computation on hardware everyone already has. The work is grounded in the mismatch between tensor cores built for narrow math and the wide-integer arithmetic of schemes like CKKS.
Catalog record
| YEAR | 2026 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-22229