AI Infra Wiki

标签: now

此标签下有7条笔记。

  • 2026年10月06日

    Fovea: Physical-Implication-Aware Wafer-Scale DSE

    • wse
    • now
    • network-on-wafer
    • noc
    • interconnect
    • chiplet
    • packaging
    • architecture
    • training
    • llm
    • optimization
  • 2026年10月06日

    Network Design for Wafer-Scale Systems with Wafer-on-Wafer Hybrid Bonding

    • wse
    • now
    • network-on-wafer
    • wafer-on-wafer
    • hybrid-bonding
    • noc
    • interconnect
    • chiplet
    • mesh
    • routing
    • 3d
    • packaging
    • training
    • llm
    • architecture
  • 2026年10月06日

    Mozart: Modularized and Efficient MoE Training on 3.5D Wafer-Scale Chiplet Architectures

    • wse
    • chiplet
    • moe
    • training
    • hybrid-bonding
    • 3d
    • noc
    • interconnect
    • parallelism
    • communication
    • llm
    • architecture
    • now
    • network-on-wafer
  • 2026年9月10日

    Network-on-Wafer

    • now
    • network-on-wafer
    • wafer-on-wafer
    • wse
    • noc
    • interconnect
    • hybrid-bonding
    • chiplet
    • mesh
    • 3d
    • packaging
    • fabric
    • architecture
  • 2026年9月10日

    WaferTrans: IOMMU-free Distributed VA Translation for Wafer-scale GPUs

    • wse
    • now
    • network-on-wafer
    • gpu
    • memory
    • interconnect
    • mesh
    • architecture
    • scale-up
    • fabric
    • virtualization
  • 2026年8月27日

    Thermal Tuning Overhead in Wafer-Scale Optical Interconnects for LLM MoE Training

    • now
    • network-on-wafer
    • wse
    • photonic
    • optical
    • cpo
    • wdm
    • interconnect
    • hybrid-bonding
    • 3d
    • packaging
    • moe
    • training
    • llm
    • architecture
    • fabric
    • scale-up
  • 2026年8月26日

    Hot Chips 2026: Cerebras CS-4

    • cerebras
    • wse
    • now
    • network-on-wafer
    • inference
    • rack
    • sram
    • 3d
    • interconnect
    • architecture

Created with Quartz v4.5.1 © 2026

  • Source Wiki
  • Quartz