HAWK: Heterogeneity-Aware KV Cache Transfer for Efficient GPU-NPU Phase-Disaggregated LLM Inference

arXiv:2610.29873 · cs.AR, cs.DC · Submitted 2026-10-21 · Read on arXiv