# R02: ReDASH: Fast and efficient Scaling in Arithmetic Garbled Circuits for Secure Outsourced Inference

ReDASH generalizes residue-number-system scaling and uses ScaleQuant+ to reduce encoded-inference overhead. PLLM status: reference primitive only.

[View canonical HTML](https://pllm.run/research/papers/r02-redash/)

Document ID: `pllm.docs.research.papers.r02-redash`  
Release: `0.1.0`

**Status:** `reference primitive only`
**Year:** 2025
**Authors:** Felix Maurer, Jonas Sander, Thomas Eisenbarth
**External source:** [ReDASH: Fast and efficient Scaling in Arithmetic Garbled Circuits for Secure Outsourced Inference](https://arxiv.org/html/2506.14489v1)

## Source contribution

ReDASH generalizes residue-number-system scaling and uses ScaleQuant+ to reduce encoded-inference overhead.

## PLLM perspective

Implement signed base extension, the published scaling gadget and its quantization policy as separately versioned components. Expose exact and approximate semantics explicitly.

Only researched previously; earlier CRT kernels did not implement full ReDASH scaling.

PLLM has related clean-room reference primitives only. They are not a paper reproduction, reviewed protected runtime, or complete-model result.

The source's boundary remains binding: Preserve the specified sign/scaling assumptions. Modular inversion is not ordinary fixed-point division; per-residue ReLU is not global integer ReLU.

## PLLM modules and documentation

Target module identifiers: `convert.rns_base_extension`, `gate.scale`, `numeric.scalequant_plus`. These identifiers describe scoped work unless this page explicitly links an implementation source. Dependencies: R01, R03.

- [Research backlog](/research/backlog/)
- [Reimplementation backlog](/research/backlog/)
- [Clean-room workflow](/research/clean-room/)
- [Arithmetic garbling documentation](/sdk/pipeline/protocols/garbling/arithmetic/)
- [Reference garbling source](https://github.com/blairhudson/pllm/blob/main/crates/pllm-garble/src/lib.rs)

## Backlog

This source is assigned to reimplementation priority 3, **Arithmetic garbling foundation**. Its source-registry order is 2. Follow current gates and unfinished work in the [research backlog](/research/backlog/).