Announcement

Collapse
No announcement yet.

It Turns Out RISC-V Hardware So Far Isn't Entirely Open-Source

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • It Turns Out RISC-V Hardware So Far Isn't Entirely Open-Source

    Phoronix: It Turns Out RISC-V Hardware So Far Isn't Entirely Open-Source

    While free software/hardware advocates have been ecstatic about the RISC-V open-source, royalty-free processor architecture, hardware so far hasn't been as open as desired...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Turns out making the analogue side of a DRAM controller is much like performing witchcraft. SiFive hasn't had the time or manpower to do that proper. On the other hand, from the digital world it's a fairly dumb component, so the ROI of reinventing this wheel is minimal. Instead of making such a big investment they decided to allocate their resources on processor architecture and production, Unfortunately this means their hands are now tied to someone like Cadence for the DRAM controller.
    An understandable move which you expect to tick off the purest of purest GNU fanatics who don't like half-empty glasses. Most people though are impressed the glass is already half full.
    Last edited by RSpliet; 24 June 2018, 06:39 PM.

    Comment


    • #3
      Maybe I'm missing something, but why was this a surprise in the first place ?

      The RISC-V core is open source but that does not include any of the other blocks required to make a CPU/SOC, like the DRAM controller in this case. AFAIK those still have to rely on proprietary designs with the associated restrictions. I imagine that will change over time though.

      Originally posted by RSpliet View Post
      Turns out making the analogue side of a DRAM controller is much like performing witchcraft.
      Yeah, witchcraft is a pretty good description... although at least DDR3 at stock clocks is a bit less witchy than some these days.

      The analog bits are also much more fab-process-specific than the core.
      Last edited by bridgman; 24 June 2018, 06:57 PM.
      Test signature

      Comment


      • #4
        Yeah, it's not going to come all at once. The parts that are RISC-V itself are certainly available in an open source form.

        Comment


        • #5
          Originally posted by RSpliet View Post
          Turns out making the analogue side of a DRAM controller is much like performing witchcraft.
          What is the "analogue side" of a DRAM controller?

          Comment


          • #6
            Originally posted by RSpliet View Post
            you expect to tick off the purest of purest GNU fanatics who don't like half-empty glasses. Most people though are impressed the glass is already half full.
            It's been said that there are 3 types of people:

            Those that see the glass as half full.

            Those that see the glass as half empty.

            Those that realize the first 2 are missing the point, the glass is refillable.

            Comment


            • #7
              Originally posted by starshipeleven View Post
              What is the "analogue side" of a DRAM controller?
              The connection between DRAM and DRAM controller requires a surprisingly large amount of tricks to maintain perfect signal integrity at high speed. Think phase-shifted clocks to shift the data sample time at sub-clock granularity, impedance matching for reduced noise and per-line delays to align their data valid windows... if I knew everything about it I'd be out there creating an open source memory controller rather than complaining about how difficult it was ;-). Related, there's this blog post from an open-source video camera implementation effort that illustrates the complexity of developing a DRAM controller for certain FPGAs, where Xilinx had taken care of the analogue DRAM interface and Elphel's design "only" had to configure it correctly for DDR3 running at a mere 400MHz.
              Last edited by RSpliet; 24 June 2018, 10:27 PM.

              Comment


              • #8
                This story, in one fell swoop, explains why Linux on the desktop never has, and never will, made any real inroads into breaking the Windows monopoly that exists in personal computers.

                There is this idiotic ideology that has a strangle hold on the open source community that holds that unless any given part is open source then it should not be part of the overall system.

                People bitch about proprietary blobs, drivers, firmware, now we have people bitching about hardware that is also not 100% "open source". I want to ask anyone that has a problem with any proprietary technology in a PC, what they hell they are smoking.

                They expect someone to study hard in school, spend the money and time to go to college, get a degree, have significant student loan debt and then work for free coding software and designing and building hardware. They get all pissed off if some company, that made a substantial monetary and time investment is hiring people to design and build hardware to then turn around and make their designs "open" so that the whole world can copy them.

                I guarantee you that if any of these people showed up for work one day and their company told them they would have to work for free, they would throw a fit.

                They expect software to not only be free, but also freely distributed as they see fit, they expect the hardware vendors to not have any patents on their designs, they expect everything to be "open". I'm wondering do these people also expect a restaurant or a supermarket to give away their goods and services for free as well? How do they feel about a chef that creates a dish that people love but that refuses to release his recipes so that he can keep his competitive advantage.

                This is coming from a devout Linux user, I'm typing this on a system running Solus, the open source community needs to collectively grow up and stop acting like a bunch of hippie, barely in their 20's, fledgling Stalinists.

                Comment


                • #9
                  LowRISC are making a fully open RISC-V board. They are actually pretty far along, but not quite there yet. Also, despite having a full-featured RISC-V CPU it won't win any benchmarks.

                  Comment


                  • #10
                    Originally posted by bridgman View Post
                    The RISC-V core is open source but that does not include any of the other blocks required to make a CPU/SOC,
                    Core or ISA? The quote from the article suggests only the ISA is open, which I think is also true of SPARC, nowadays (and POWER, as noted).

                    You can even supposedly get open source implementations:

                    Comment

                    Working...
                    X