Latest

• FastDecode: High-Throughput GPU-Efficient LLM Serving using Heterogeneous Pipelines
• POSTER: Pattern-Aware Sparse Communication for Scalable Recommendation Model Training
• SmartMoE
• SmartMoE: Efficiently Training Sparsely-Activated Models through Combining Offline and Online Parallelization
• FasterMoE
• BaGuaLu: targeting brain scale pretrained models with over 37 million cores
• Efficiently emulating high-bitwidth computation with low-bitwidth hardware
• FasterMoE: modeling and optimizing training of large-scale dynamic pre-trained models
• AIPerf
• FastMoE
• Critique of \"Planetary Normal Mode Computation: Parallel Algorithms, Performance, and Reproducibility\" by SCC Team From Tsinghua University
• FastMoE: A Fast Mixture-of-Expert Training System
• Student Cluster Competition 2018, Team Tsinghua University: Reproducing performance of multi-physics simulations of the Tsunamigenic 2004 Sumatra megathrust earthquake on the Intel Skylake Architecture