The Structure of In-Place Space-Bounded Computation
Abstract
In the standard model of computing multi-output functions in logspace (), we are given a read-only tape holding and a logarithmic length worktape, and must print to a dedicated write-only tape. However, there has been extensive work (both in theory and in practice) on algorithms that transform into in-place on a single read-write tape with limited (in our case ) additional workspace. We say if can be computed in this model. We initiate the study of in-place computation from a structural complexity perspective, proving upper and lower bounds on the power of . We show the following: i) Unconditionally, . ii) The problems of integer multiplication and evaluating circuits lie outside under cryptographic assumptions. However, evaluating circuits can be done in . iii) We have Consequently, proving would imply . We also consider the analogous catalytic class (), where the in-place algorithm has a large additional worktape tape that it must reset at the end of the computation. We give algorithms for matrix multiplication and inversion over polynomial-sized finite fields. We furthermore use our results and techniques to show two novel barriers to proving . First, we show that any proof of must be non-relativizing, by giving an oracle relative to which . Second, we identify a search problem in but not known to be in .
Keywords
Cite
@article{arxiv.2510.12005,
title = {The Structure of In-Place Space-Bounded Computation},
author = {James Cook and Surendra Ghentiyala and Ian Mertz and Edward Pyne and Nathan S. Sheffield},
journal= {arXiv preprint arXiv:2510.12005},
year = {2025}
}
Comments
43 pages