Hot Chips 2026: Applying High Bandwidth Flash (HBF)

(chipsandcheese.com)

21 points | by ksec an hour ago ago

4 comments

  • rando1234 34 minutes ago

    What will be the durability/lifetime properties of this technology in comparison to traditional DRAM?

    • LogTrim 32 minutes ago

      Likely much worse write endurance than DRAM, since it’s still flash underneath. The saving grace is that model weights are mostly read-heavy, so endurance may matter less than it sounds.

  • ksec an hour ago

    I had to double check those figures on Sk Hynix office web site [1], and it is not a typo or wrong capital "B".

    It really is 3TB per second.

    I literally paused for 5 min and thought how is this even possible.

    [1] https://news.skhynix.com/en/hbf-at-fms-2026/

    • amluto 29 minutes ago

      This is almost entirely dominated by the read circuitry and the data path: it’s still taking 1/6 of a second to read the whole chip, which means that the flash cells aren’t working hard at all. (And that pitting the full weights of a dense model on these chips while using anywhere near all the capacity is a nonstarter if you intent to stream the weights as you run inference.)