Microsoft

LeanAttention: Hardware-Aware Scalable Attention Mechanism for the Decode-Phase of Transformers

Transformer-based models have emerged as one of the most widely used architectures for natural language processing, natural language generation, and image generation. The size of the state-of-the-art ...
IEEE

Implementation of LDPC Decoder using Barrel Shifter

Abstract: The research work explores the implementation of an LDPC decoder utilizing a barrel shifter architecture tailored for processing 128-bit data. LDPC codes ...
IEEE

A Novel Implementation of Ternary Decoder Using CMOS DPL Binary Gates

Abstract: This paper proposes a novel implementation of Ternary Decoder using CMOS DPL (Double Pass Logic) Binary logic gates, in digital CMOS technology. The physical design of the circuits is ...
GitHub

Kernel Library for LLM Serving

FlashInfer is a library and kernel generator for Large Language Models that provides high-performance implementation of LLM GPU kernels such as FlashAttention, SparseAttention, PageAttention, Sampling ...