Hey everyone,

I’ve noticed this community is a bit barren. I’d like to add some posts myself but from a more ignorant POV.

It has come to my attention that there is not nearly enough software torture in my life. Hearing how companies are starting to add barriers to so-called “Translation layers” and other software pieces that give functionality to the underlying hardware…

What types of projects does the open source community need to pursue with respect to RISC-V drivers, firmware, or other necessary pieces to somewhat guarantee that the processor is compatible with GPUs, RAM, I/O, and other low level functions that hardware components perform?

Where should I start in my RISC-V journey? Is there a beginner computer or developer kit that I can purchase that won’t incinerate my wallet? I don’t expect to game or train LLMs off of a RISC-V-based computer, but am curious as to what a fully free hardware setup would consist of.

Love the idea of RISC-V!

  • phanto@lemmy.ca
    link
    fedilink
    arrow-up
    5
    ·
    9 months ago

    I have a Vision Five board. It cost me 80$ on super early bird Kickstarter, back when it first launched. I’ll be honest, it’s collecting dust. I keep meaning to get it set up, but I am too distracted. One day…

    • gronjo45@lemm.eeOP
      link
      fedilink
      arrow-up
      1
      ·
      9 months ago

      Maybe if we got a Matrix chat going for this server that the dust could be brushed off… Perhaps it’s time :D

  • just some guy
    link
    fedilink
    arrow-up
    5
    ·
    9 months ago

    Pine64 has some risc-v stuff. They’re usually pretty fairly priced from my experience

    • gronjo45@lemm.eeOP
      link
      fedilink
      arrow-up
      1
      ·
      9 months ago

      The Pinetab-V and Star64 look interesting for a tablet and SBC on the RISC-V architecture… What Linux distribution would one have for options on that processor architecture? Debian sounds appealing there…

      • just some guy
        link
        fedilink
        arrow-up
        1
        ·
        edit-2
        9 months ago

        I know there’s an Armbian image (probably Debian based) and a few others for the star64, while iirc the default os for the PineTab v is a yocto build. I’m sure it would be possible to make a Debian img for either.

    • gronjo45@lemm.eeOP
      link
      fedilink
      arrow-up
      2
      ·
      9 months ago

      Especially with all the hoot around LoRA ML models and AI training methods that don’t require a nuclear reactor and loads of proprietary offloading servers. Maybe we could get a Matrix chat going for those interested in things like kernel development, and work on some educational projects?

    • gronjo45@lemm.eeOP
      link
      fedilink
      arrow-up
      1
      ·
      9 months ago

      Exactly what I asked in a previous comment! This looks awesome, but what about Pine64’s hardware?

  • j4k3@lemmy.world
    link
    fedilink
    arrow-up
    2
    ·
    9 months ago

    I only continue to monitor reports by DJ Ware on YT. He is Bell Labs alumni and talks proxmox, distros, and stuff like the state of RISC V. Someone had announced an entire data center running RISC V within the last year. I’m clueless on anything deeper though.

    • gronjo45@lemm.eeOP
      link
      fedilink
      arrow-up
      2
      ·
      9 months ago

      I love DJ Ware after your recommendation! HPCs running RISC-V would be quite interesting…

      • j4k3@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        9 months ago

        Any time I hear someone is Bell Labs alumni, it gets my attention. I know of DJ Ware, Curious Marc, and The Signal Path. Marc is doing a bunch of stuff for the Computer History Museum and reverse engineering Apollo, and Shahriar is a gigahertz radio wizard for Bell Labs currently. Both links are to their personal blogs, but they are on YT as well.

        I got into AI and took note of Yann LeCunn as the head of Meta AI. He is also Bell Labs alumni and one of the main forces behind offline open source AI.

        If you or anyone else knows of Bell Labs alumni posting publicly facing content. Don’t hesitate to let me know. That credential tends to mean a whole lot in the grand scheme.