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

## Episode Summary In this episode, we cover: - **HAWK: Heterogeneity-Aware KV Cache Transfer for Efficient GPU-NPU Phase-Disaggregated LLM Inference** (arXiv) - **Accelerating CPU Speculative Decoding via Unified Graph Execution and Transactional KV Isolation** (arXiv) - **Why Chip Design AI Benchmarks Must Go Beyond Functional Correctness** (eetimes) - **SpacemiT V100: New RISC-V Server Processor Combines 46 Cores, DDR5 and PCIe 5.0 - igor´sLAB** (google_riscv) - **Alibaba unveils automotive RISC-V processor IP and roadmap for self-driving and robotics - TechNode** (google_riscv) --- *Sponsored by Ada, Ago Consulting, and Zen Semiconductor*