Meme transcription:

[Guy looking happy:] Hey, there’s a Linux version of that program I like

[Guy looking disappointed:] Closed Source, no ARM support

  • ZkhqrD5o@lemmy.world
    151·
    3 days ago

    My Little Rant on Risc V and Arm vs x86, because I have an opportunity to dump this here and get it off out of my system.

    RISC systems will maybe perhaps take over market share from x86 in the mobile laptop space, but essentially there’s no point in anything else. If I remember correctly, RISC came before CISC in computing, and many old mainframes were RISC. But then people started thinking: “What if we can do multiple operations in a single instruction?” And therefore CISC was born and he did wonders for performance.

    Yes, the reduced instructions are very nice for battery life. Who doesn’t like good battery life? People who like performance, that’s for sure. So if you run a programme that is programmed for CISC primarily, and then you just change the compiling target to a RISC system, then you will basically use the same battery life, but with worse performance than just using a CISC system, Since multiple clocks now need to do something that has happened in a single clock on CISC. I fully understand that having a monopoly on computing hardware is very bad. I don’t get the hype around arming normal computers because it will just shift the monopoly from one to another, the harm to innovation remains. Risc V is interesting because it would break the monopoly, but the problem is it uses a pushover license. So companies would reap all of the benefits for developing a proprietary risk system, and everyone who likes to compete is free to use the reduced, almost unusable base spec. I mean, compare the BSD kernel to the Linux kernel. It’s nowhere close. So with that being said, I think x86 in public domain would be the nicest thing to happen. Thank you for listening to my useless TED talk.

    Edit: Thanks for the interesting replies, people! Time will tell what will have happened, so let’s find out together.

    • Treczoks@lemmy.world
      1·
      2 days ago

      If you want to replace an x86 based machine with a RISC-V based one, you are in for a bad surprize: speed, or, more precisely, the absence if it.

      While a good ARM core can already be useful in a PC, the RISC-V is still far behind in performance.

      Funny, though, Trump might unintentionally fix this by driving China to use non-X86, non-ARM platforms, and here, RISC-V is the best choice. When the Chinese throw their rather substantial R&D power behind this architecture, it might actually getting somewhere.

    • thepreciousboar@lemm.ee
      12·
      3 days ago

      Just because a CISC will run multiple instructions in a single clock cycle, it doesn’t automatically make it faster. Complex instructions means complex decode logic, that makes the execution slower, even at the same clock cycle. A modern intel CPU has something like 20+ stages of pipelining, while ARM has 3-5 stages, that makes the execution more energy efficient and more powerful. Also superscalar RISC architectures exist, so RISC can also execute more instructions at a time, and in less time.

      Lastly, modern x86_64 look like CISC, but are actually RISC under the hood, the single instruction is just a pseudo-instruction divided in multiple simpler instructions. I don’t believe thay makes it much more efficient.

    • exu@feditown.comEnglish
      9·
      3 days ago

      RISC vs CISC doesn’t really matter. Both have heavily borrowed from each other. The big differences are design goals, x86 processors are targeting higher power processing with few very fast cores while Arm and Risc V mostly target embedded and low power computing or a huge number of smaller cores.

    • lime!@feddit.nuEnglish
      6·
      3 days ago

      SIMD and CISC are not tied together. ARM has SIMD. RISC-V has vector extensions like old mainframes. they can both do SIMD and combined instructions jush fine. they just don’t have dedicated circuitry for them.