5 papers
Correlation-Aware Structured Pruning for Large Language Models
Sicheng Xu, Hao Shi, Wei Zhang +6
Structured pruning is a promising approach for reducing the substantial inference costs of Large Language Models (LLMs) while maintaining hardware efficiency. Many existing methods…
Tight Lower Bounds on The Single-Error Detection Threshold for Analog Error-Correcting Codes
Zhengyi Jiang, Wenhao Liu, Zhongyi Huang +3
Analog error-correcting codes (Analog ECCs) for approximate vector-matrix multiplication have been extensively studied as means to achieve fault-tolerant in-memory computation. The…
Grid-like Error-Correcting Codes for Matrix Multiplication with Better Correcting Capability
Hao Shi, Zhengyi Jiang, Zhongyi Huang +3
Matrix multiplication over the real field constitutes a foundational operation in the training of deep learning models, serving as a computational cornerstone for both forward and…
Error Correction Decoding Algorithms of RS Codes Based on An Earlier Termination Algorithm to Find The Error Locator Polynomial
Zhengyi Jiang, Hao Shi, Zhongyi Huang +4
Reed-Solomon (RS) codes are widely used to correct errors in storage systems. Finding the error locator polynomial is one of the key steps in the error correction procedure of RS c…
Set Transformation: Trade-off Between Repair Bandwidth and Sub-packetization
Hao Shi, Zhengyi Jiang, Zhongyi Huang +3
Maximum distance separable (MDS) codes facilitate the achievement of elevated levels of fault tolerance in storage systems while incurring minimal redundancy overhead. Reed-Solomon…