Google released TPU Sync, an open-source C++ library that moves KV caches and weights across TPU HBM, host DRAM, storage and network. Google reports over 90% lower KV transfer latency on 235B to 480B MoE models.
This article has been indexed from MarkTechPost
Read the original article:
Google Releases TPU Sync: Open-Source KV Cache Transfer for TPUs