Proposal: Accelerated State-Vector Compression using DPX-based Baire Metric indexing

[StanByriukov02]

The cuQuantum team, I'm working on state compression in large-scale quantum simulations.

I've discovered that using the Baire Metric, implemented in DPX instructions, allows for deterministic amplitude indexing with linear complexity.

Unlike standard hashing methods, this approach preserves the hierarchical structure of the data, which is critical for tensor networks.

I've implemented Protocol Omega—a bridge between DPX dynamic programming and tensor logic. The results show a significant reduction in computational noise and thermal footprint.

I'd be happy to share proofs (receipts) and discuss applicability in future SDK versions.
CTDR_public_pack_20251219.zip

[yangcal]

Thanks for the interest and suggestion. For security reasons, we can’t download, open, or execute attachments (including zip files) from GitHub Discussions.

If you’d like this considered, please repost the proposal without attachments and include the technical details directly in the thread (or link to a public paper/repo): a clear problem statement, algorithm description, complexity analysis, minimal reproducible example, benchmark methodology/results, and how it integrates with cuQuantum APIs.

Otherwise we’ll close this discussion as not actionable.