61 pointsby ingve3 hours ago8 comments
  • conspan hour ago
    So you basically dispose of cache for the memory region used? I wonder what the offsets of the cache misses is going to be in practice (the article addresses it but there is no solution/impact given by samsung).
    • manmal41 minutes ago
      I guess you also get very high bandwidth that way? I‘m not sure that would come for free though.
    • PunchyHamster17 minutes ago
      if your working set fits in cache PIM is irrelevant
      • yvdriess7 minutes ago
        Caches are important for cpu core performance even when the the working set doesn't fit.
  • reliabilityguy2 hours ago
    Interesting that Samsung still pursues PIM. IIRC they had a paper in ISCA21 or 22 where they showed HBM2 module with PIM, which back then impressed me quite a lot.

    That being said, I am not sure what’s the killer application for this technology, and without such application adoption is unlikely.

    • xyzzy123an hour ago
      As I understand it, the killer app is llms. You could run MACs directly in RAM, offloading a lot of work from CPU and cutting down on insane (external) memory bandwidth required.

      Imagine (this is a fantasy pitch but potentially achievable for some use cases) wanting to run a larger llm and all you have to do is buy more RAM so it fits.

      • amelius9 minutes ago
        You: "AI, please write me $COOL_APP."

        AI: "Sorry, all the hardware is made for running AI."

      • embedding-shapean hour ago
        > Imagine (this is a fantasy pitch but potentially achievable for some use cases) wanting to run a larger llm and all you have to do is buy more RAM so it fits.

        Isn't this how it works today already? Granted you wanted to run it on RAM rather than VRAM.

        • petua minute ago
          Yes, but running out of RAM is impractical due to low memory bandwidth.

          According to article/Samsung RAM dies inside can provide way more bandwidth, they're limited by external interface / bus width:

          > Together, they can utilize the chip’s internal bandwidth across all 16 banks, which comes out to 614 GB/s. For comparison, regular DRAM accesses can hit two banks in parallel and max out at 76.8 GB/s.

          And that's just for single 64-bit IC. So way faster and more power efficient.

    • netfortiusan hour ago
      > That being said, I am not sure what's the killer application for this technology...

      Build it, and they will come ;)

  • userbinatoran hour ago
    In-memory computation was already possible with regular DRAM: https://news.ycombinator.com/item?id=22712811

    Add a new set of CPU instructions like “rep macb”

    ...and it's been long enough now, that I can say there was an effort to implement this on standard x86 memory controllers and have the existing string instructions do so, back in the days of SDR SDRAM, but the tradeoffs weren't (yet) in favour.

  • saejox43 minutes ago
    If we could buy a 64gb stick and run a 32b model with 30tps on it. This could sell
  • Torkel26 minutes ago
    [dead]
  • LogTriman hour ago
    [flagged]
  • sehw2 hours ago
    [dead]
  • johnnyApplePRNGan hour ago
    [flagged]