Skip to main content

ConferenceMay, 2026Published

I/O-Aware PIM Acceleration for Long-Sequence LLM Inference with Hybrid Sparse Attention

Authors: X. Lu, L. Hu, H. Huang, P. Jiang, X.-H. Sun
Venue: Proceedings of the 40th IEEE International Parallel and Distributed Processing Symposium (IPDPS 2026)
Date: May, 2026
Type: Conference
DOI: 10.1109/IPDPS65963.2026.00046 (opens in a new tab)

Knowledge Graph Connections

Core Contribution To

UniMCC

Core Contribution To

Memory Access Pattern Obfuscation