• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI
    Report

    HBM controllers and PHYs estimated to use about 4% of Feynman's GPU die, versus roughly 16% on Rubin

    SemiAnalysis says the memory controller moves onto the HBM base die, while Nvidia's compact NV-HBI link replaces the standard PHY.

    SE
    1 Source, 3h ago, first seen 3h ago

    TLDR

    SemiAnalysis estimates that HBM controllers and PHYs use roughly 16% of Rubin's GPU die, falling to about 4% on Feynman with NVHBM. It says the memory controller moves to the HBM base die and Nvidia's NV-HBI link replaces the standard PHY. Nvidia claims up to 25% more die area for compute, up to 30% more bandwidth and 15% lower HBM power than standard HBM4E.

    Combined views

    14.7K

    1 Source, first seen 3h ago

    Combined views

    14.7K

    1 Source, first seen 3h ago

    138 likes
    138 likes
    11 comments
    53 saves
    15 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Featured Source
    11 comments
    53 saves
    15 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    1 Source

    @SemiAnalysis_🚨Custom HBM frees up die space for compute🚨 On Rubin, HBM controllers and PHYs take up roughly 16% of the GPU die. On Feynman with NVHBM, we estimate that falls to ~4%. How: the memory controller moves off the XPU and onto the HBM base die, and the wide standard PHY is replaced by Nvidia's compact NV-HBI die-to-die link. Samsung's custom HBM4 interface is ~60% smaller than its standard PHY. Nvidia claims up to 25% more die area for compute, up to 30% more bandwidth, and 15% lower HBM power vs standard HBM4E. (1/2)🧵7h

    1 Source

    @SemiAnalysis_🚨Custom HBM frees up die space for compute🚨 On Rubin, HBM controllers and PHYs take up roughly 16% of the GPU die. On Feynman with NVHBM, we estimate that falls to ~4%. How: the memory controller moves off the XPU and onto the HBM base die, and the wide standard PHY is replaced by Nvidia's compact NV-HBI die-to-die link. Samsung's custom HBM4 interface is ~60% smaller than its standard PHY. Nvidia claims up to 25% more die area for compute, up to 30% more bandwidth, and 15% lower HBM power vs standard HBM4E. (1/2)🧵7h