1

We propose OctoPipe, a novel pipeline parallelism system that reduces pipeline bubbles on heterogeneous models by co-optimizing partitioning, placement, and scheduling, achieving 1.22-2.14x throughput improvement over Megatron-LM.

We propose UniTG, the first unified system that fuses the LM and GNN phases of textual graph learning into a single end-to-end procedure, reducing learning makespan by up to 17.3x without compromising model quality.

Traditional ID-based recommender systems often struggle with cold-start and generalization challenges. Multimodal recommendation …

Batch inference for agentic workloads stresses the GPU key-value (KV) cache in a sustained and cumulative manner, often causing severe …

GPU checkpointing and restore promises to enable emerging tasks, such as deep learning, to benefit from functionalities like task …

Offloading large language models (LLMs) state to host memory during inference promises to reduce operational costs by supporting larger …

Hybrid parallelism underpins large-scale LLM training across tens of thousands of GPUs. At such scale, hardware failures on individual …

Large Language Models (LLMs) have presented impressive performance across several transformative tasks. However, it is non-trivial to …

Hyperparameter tuning is an essential step in deep learning model development that provides better model performance at the cost of …

With the proliferation of deep learning, there exists a strong need to efficiently operate GPU clusters for deep learning production in …