Fast Evaluation of Truncated Neumann Series by Low-Product Radix Kernels
Piyush Sao · 2026
Read Fast Evaluation of Truncated Neumann Series by Low-Product Radix Kernels by Piyush Sao (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-2602-11843.txt.
Description
Truncated Neumann series approximate matrix inverses and precondition polynomials, but in dense settings matrix-matrix products dominate their cost. Naive evaluation needs one product per term, while splitting methods cut that to logarithmic counts. Repeated squaring uses about twice the log count, so further gains need higher-radix kernels that extend the series by several terms per update. Piyush Sao constructs such kernels beyond the previously known radix-five case.
For radix nine the author derives an exact three-product kernel with rational coefficients, the first exact construction beyond radix five, yielding a 21 percent product reduction from repeated squaring. For radix fifteen, numerical optimization gives a four-product kernel matching the target through degree fourteen, with spillover at higher degrees handled by a new residual-based framework. That framework attains the best known asymptotic rate, and experiments support the predicted savings and runtime trends.
Catalog record
| CREATOR | Piyush Sao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-11843