David Shu-Yu Wu

Multiplying Not-So-Big Integers with FFTs

Sat, 04 Jul 2026 00:00:00 +0000

This first-author research project, Finding and Pushing the Crossover on Large Modern OoO ARM CPUs, studies where NTT-based multiplication becomes practical for not-so-big integer sizes. The paper is available as and was accepted for presentation at the CHES 2026 Poster Session.

The project develops two generations of constant-time NTT-based multiplication for Cortex-A76. The second generation targets smaller operand sizes through tighter range bounds, redesigned 256/512-point NTT organization, fused reconstruction, and shorter carry and instruction-dependency chains.

Key Features

NEON-optimized NTT kernels: Developed constant-time fused NTT kernels in AArch64 assembly using ARM NEON instructions.
Multiplication pipeline: Combined Barrett and Montgomery arithmetic, Good’s trick, CRT reconstruction, and chunking and dechunking.
Crossover result: The second-generation two-unknown-input multiplier outperforms target-tuned variable-time GMP by 6.25% at 5,632 bits and 16.66% at 8,448 bits, placing the observed crossover at or below 5,632 bits.
Montgomery multiplication: The one-known-input implementation improves over OpenSSL’s Neon path by 5.19%-44.40% across the evaluated 5,120-10,240-bit sizes.
OpenSSL integration: End-to-end RSA verification and encryption improve by 15.98%-51.58% across RSA-5120, RSA-6144, RSA-8192, and RSA-10240.
Formal verification: Verified optimized assembly kernels with CryptoLine for algebraic correctness, range safety, and equivalence after Slothy scheduling, then used HOL Light to compose higher-level multiplication specifications.

Generative AI for High-Fidelity Side-Channel Trace Simulation

Fri, 03 Jul 2026 00:00:00 +0000

This research began through the NTU Global Connect Fellowship Program at Nanyang Technological University and continues as a remote collaboration, supervised by Gwee Bah Hwee and Juncheng Chen. A manuscript is in preparation.

The project studies how attack-relevant leakage emerges during diffusion-based side-channel trace generation and whether cryptanalytic utility develops at the same rate as waveform and distribution fidelity.

Experiments and Findings

Datasets: Developed conditional DDPM experiments on controlled synthetic traces and four measured AES datasets.
Reverse-trajectory evaluation: Evaluated predicted-clean traces across all 400 reverse-diffusion stages rather than only the final samples.
Attack-based evaluation: Measured CPA-based key-recovery complexity throughout the reverse trajectory.
Fidelity evaluation: Combined waveform and distribution distances with nearest-neighbor likelihood scoring.
Key observation: Found that attack-relevant leakage can emerge substantially earlier than full distributional fidelity.
Masked traces: Matching first-order marginals alone was insufficient to reproduce the cross-share dependence required for second-order attacks.

Cryptography Engineering on Cortex-M4

Thu, 02 Jul 2026 00:00:00 +0000

This project is the final project for Cryptography Engineering (2025 Fall) at National Taiwan University, lectured by Matthias J. Kannwischer. It focuses on aggressive performance optimization of two modern cryptographic primitives on resource-constrained embedded systems.

The project implements highly optimized versions of:

ML-KEM (FIPS 203): NIST’s lattice-based post-quantum key encapsulation mechanism
X25519 (ECDH25519): Elliptic Curve Diffie-Hellman key exchange

Target Platform: ARM Cortex-M4 microcontroller
Constraints: Constant-time execution, strict code-size budgets for NTT routines

Key Optimizations

X25519 (ECDH25519)

Finite Field Arithmetic (mod p = 2^255 - 19): Hand-optimized assembly for fe25519_mul (field multiplication) and fe25519_sqr (field squaring), achieving 81-85% cycle reduction.
Constant-Time Scalar Multiplication: No secret-dependent branches or table indexing to prevent timing side-channel attacks.
Fixed-Base Acceleration: Window method with precomputed tables using constant-time selection, reducing fixed-base scalar multiplication by up to 94% against the course baseline.

ML-KEM (FIPS 203)

NTT/iNTT Kernels: Optimized Number-Theoretic Transform operations with reduced memory traffic, layer fusion, and loop unrolling, achieving 2.85-3.79x speedup.
Polynomial Arithmetic (mod q = 3329): Leveraged ARMv7E-M DSP instructions (SIMD 16-bit operations) for packed polynomial addition and subtraction, achieving 2-2.2x speedup.
Code-Size Budgeting: NTT routines implemented under strict per-function size constraints while maintaining performance.
Course-Baseline Result: Reduced the cycle count of ML-KEM poly_ntt by up to 62% under the same benchmarking setup.

This project demonstrates deep understanding of low-level optimization, cryptographic algorithm engineering, and secure implementation practices on embedded ARM platforms.

Fuzzy Counting with Probabilistic Data Structures

Wed, 01 Jul 2026 00:00:00 +0000

This Advanced Data Structures course project explores typo-tolerant fuzzy counting with compact probabilistic data structures.

Key Features

Fuzzy matching: Implemented typo-tolerant matchers using Bloom filter segmentation and Cuckoo filters.
Cardinality estimation: Implemented and compared HyperLogLog and HyperLogLog-LDP.
Evaluation: Measured accuracy-space tradeoffs over synthetic typo datasets under different sketch parameters and segmentation strategies.
Reusable tooling: Exposed the implementation as a Python library backed by the C++ experiments.

You can view the complete source code and the full academic report from the links below:

Algorithm Analysis & Data Structure Experiments

Sun, 01 Jun 2025 00:00:00 +0000

This repository is a collection of in-depth academic projects for an Advanced Data Structures course, focusing on the implementation and comparative analysis of complex algorithms.

Key Components

BST Performance Analysis: Implemented and conducted experiments on various Binary Search Trees, including Splay Trees and Tango Trees, to analyze their performance bounds. Includes a full academic report.
Streaming Algorithms Comparison: Implemented and benchmarked classic streaming algorithms in C++, including Morris, Flajolet-Martin (FM), and HyperLogLog (HLL), to evaluate their accuracy in cardinality estimation.
DSU Theoretical Analysis: A theoretical paper analyzing a linear-time algorithm for a special case of the Disjoint Set Union (DSU) problem.

You can view the complete source code and all academic reports from the link below:

Operating System Machine Problems (xv6)

Tue, 01 Apr 2025 00:00:00 +0000

These projects apply systems programming concepts through xv6 extensions and supporting simulations written in C.

Key Features Implemented

User-level threads: Implemented thread creation and cooperative context switching.
Kernel memory management: Built a SLAB allocator with freelist management and partial and full slab classification.
CPU scheduling: Implemented and evaluated HRRN, P-RR, DM, and EDF-CBS schedulers.
Fault-tolerant storage: Simulated RAID-1 file operations in a Docker-based development environment.

You can view all my code contributions and modifications from the link below:

Experience

Tue, 24 Oct 2023 00:00:00 +0000