Algorithm-hardware co-design

ASAP: A 28-nm Transformer Training Accelerator With Alternating Sparsity and Asymmetrical Microscaling Precision

ASAP: A 28-nm Transformer Training Accelerator With Alternating Sparsity and Asymmetrical Microscaling Precision 150 150

Abstract:

This work presents ASAP, a 28-nm transformer-training accelerator that combines N:M structured sparsity with asymmetric microscaling floating-point (MXFP) precision through a unified algorithm–hardware co-design. ASAP introduces a progressive sparsity schedule in which pruned compute resources are reassigned to increase numerical precision for important weights and activations, stabilizing optimization …

View on IEEE Xplore

SiWB: A 28-nm 800-MHz 4.2-to-14.2-Gb/s/W Configurable Multi-Core Architecture for White-Box Block Cipher With Area-Efficient Random Linear Transformation and Load-Aware Inter-Core Scheduling

SiWB: A 28-nm 800-MHz 4.2-to-14.2-Gb/s/W Configurable Multi-Core Architecture for White-Box Block Cipher With Area-Efficient Random Linear Transformation and Load-Aware Inter-Core Scheduling 150 150

Abstract:

White-box cryptography (WBC) seeks to protect secret keys (SKs) even under the white-box security model that features adversaries having full control of the execution environment. Due to the ever-growing demand for content protection under security-critical scenarios, the recent progress on WBC has been nothing short of spectacular. However, the security-prioritized …

View on IEEE Xplore