IEEE

A Streaming Dataflow Engine for Sparse Matrix-Vector Multiplication Using High-Level Synthesis

Abstract: Using high-level synthesis techniques, this paper proposes an adaptable high-performance streaming dataflow engine for sparse matrix dense vector multiplication (SpMV) suitable for embedded ...
i-SCOOP

LongCat 2.0 by Meituan pushes trillion-parameter models onto Chinese chips

LongCat 2.0 is Meituan's 1.6T-parameter MoE model with a native 1M-token context, trained entirely on domestic AI ASIC ...
IEEE

MH-SpGEMM: Efficient Sparse General Matrix-Matrix Multiplication on Modern GPUs via Masking and Hashing Cooperative Optimization

Abstract: Sparse General Matrix-Matrix Multiplication (SpGEMM) is a core operation in high-performance computing applications such as algebraic multigrid solvers, machine learning, and graph ...
GitHub

This is the official implementation of our ICML 2024 paper "MultiMax: Sparse and Multi-Modal Attention Learning".

[2026-07] MultiMax is now integrated into PaddleFleet, the large-scale distributed training framework by PaddlePaddle. Key highlights of the integration: Optimizer: Use Adam (or AdamW) for MultiMax ...