Unite.AI

124x Slower: What PyTorch DataLoader Actually Does at the Kernel Level

This article is based on findings from a kernel-level GPU trace investigation performed on a real PyTorch issue (#154318) using eBPF uprobes. Trace databases are published in the Ingero open-source ...
IEEE

Efficient KV Cache Spillover Management on Memory-Constrained GPU for LLM Inference

Abstract: The rapid growth of model parameters presents a significant challenge when deploying large generative models on GPU. Existing LLM runtime memory management solutions tend to maximize batch ...
VentureBeat

Nvidia says it can shrink LLM memory 20x without changing model weights

Nvidia researchers have introduced a new technique that dramatically reduces how much memory large language models need to track conversation history — by as much as 20x — without modifying the model ...
IEEE

Effective Memory Management and Sparse Aware Cache for Row-Wise Sparse CNNs

Abstract: Convolutional Neural Networks (CNNs) are used in several image processing tasks like image recognition and object localization. For edge applications such as drones and autonomous vehicles, ...
VentureBeat

New KV cache compaction technique cuts LLM memory 50x without accuracy loss

Enterprise AI applications that handle large documents or long-horizon tasks face a severe memory bottleneck. As the context grows longer, so does the KV cache, the area where the model’s working ...