Papers › FlexChunk: Enabling 100M×100M Out-of-Core SpMV (~1.8 min, ~1.7 GB RAM) with Near-Linear Scaling

FlexChunk: Enabling 100M×100M Out-of-Core SpMV (~1.8 min, ~1.7 GB RAM) with Near-Linear Scaling

5 Apr 2025LessWrong 2025 4archive 2025-07-28

Daniil Strizhov

Handling large-scale sparse matrices is a fundamental task in many scientific and engineering domains, yet standard in-memory approaches often hit the limitations of available RAM. This paper introduces FlexChunk, an algorithm employing a chunking strategy and disk caching for efficient sparse matrix-vector multiplication (SpMV) specifically designed for matrices exceeding available memory. Our experiments demonstrate that FlexChunk achieves near-linear O(N) time complexity and linear memory consumption scaling up to matrix sizes of 100M×100M, processing the largest case in 1 minute 47 seconds with a peak memory footprint of 1.7 GB. A comparison with the optimized SciPy library reveals a key trade-off: while SciPy offers significantly faster computation for in-memory matrices, FlexChunk excels in scenarios requiring disk I/O due to its substantially lower data loading times. The primary contribution is demonstrating FlexChunk as a scalable, memory-efficient solution enabling SpMV operations on problem sizes previously intractable due to memory constraints.

PaperPDFCode

Syntology This paper has no arXiv id in the archive, and this site joins papers to Syntology's graph by arXiv id, so this page does not say whether the graph holds it.

Code

DanielSwift1992/FlexChunk Syntology: not harvested report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

ChunkingNature-Inspired Optimization Algorithm

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections