一作ACL 2026Poster· Natural Language Processing
Deep Kernel Fusion for Transformers
Zixi Zhang · Zhiwen Mo · Yiren Zhao · Robert D. Mullins
摘要
Agentic LLM inference with long contexts is increasingly limited by memory bandwidth rather than compute. In this setting, SwiGLU MLP blocks, whose large weights exceed cache capacity, become a major yet under-optimized bottleneck in the Transformer architecture. We propose DeepFusionKernel, a de...