8 ms·

RISC-V is making moves, but it has work to do if it wants to hit the mainstream

18 points by pseudolus 2y ago

Pet_Ant 2y ago
Most people don't need more than a Chromebook has to offer. I can't imagine why this wouldn't eventually take off there and become the main CPU that people spend time on. Especially as margin tighten, the ISA-license will be will something that has to go. "Your margin is my opportunity".
- Wytwwww 2y ago
  IMHO it might be the complete opposite. Margins are low and and off the shelf ARM designs are very cheap so what's the point of making very risky investments when you can already get good enough off the shelf designs from ARM? If the licensing fee is only ~1-2% who can justify this besides Qualcomm, Apple, Nvidia, etc.? And Apple and Qualcomm only did this because they wanted to (and could afford to) design faster chips than ARM.
  - snvzz 2y ago
    There's the option of licensing a RISC-V microarchitecture, rather than an ARM one. It has the advantage that RISC-V is an open standard, and there are many companies licenses for their own RISC-V microarchitectures, rather than being stuck with licensing ARM cores from ARM.
    - Wytwwww 2y ago
      From whom? I'm not sure why would anyone who can design high-end RISC-V cores would license them to someone else instead of making the chips themselves. As ARM has demonstrated that's not a very good business model (they are barely making any money compared to some of their customers).
      - snvzz 2y ago
        >From whom? From any of the IP vendors offering them. e.g. Tenstorrent has Ascalon (already licensed by LG) and Alastor, SiFive has P670 and P870, Akeana has the 5000 series.
        Wytwwww 2y ago
        I just don't understand why would anyone who developed a competitive mid to high-end core would want to license it to anyone? What would they have to gain by doing that instead of making them themselves? Of course there are cases where the CPU is pretty much just a "cost centre" (embedded, industrial products and such) but I don't see that working in the PC/mobile/(other general purpose device) market. Doesn't really seem like a sustainable business model long-term i.e. you won't be able to generate enough revenue to compete with companies designing chips in house.
- pjmlp 2y ago
  Most people definitly need more RAM and SSD storage than most Chromebooks offer, specially when most middle range phones and tablets offer the double.
- Someone 2y ago
  For anything that sells for over $100 or so, I think not having to pay for the ISA license is nice, but the main benefit of RISC-V relative to ARM and the like is the flexibility of being allowed to add or remove features at will.
  - Wytwwww 2y ago
    That's not relevant for anyone making Chromebooks (especially low to mid end ones) though. Unless you can get an equivalent RISC-V chip significantly cheaper than an off the shelf x86/ARM one why would they do it (considering the extra costs/overhead from supporting a niche platform in addition to lost sales).
- whatevaa 2y ago
  Hardware accelerated video decode is a must-have for any light chromebooks to gain traction.
  - snvzz 2y ago
    There are many IP vendors to license video codec RTL from. RISC-V isn't a factor. Many RISC-V based SoCs already in the market have some level of hardware video codec acceleration built in.
bhouston 2y ago
I think that this article is missing the main reason why RISC-V is not yet mainstream. It is super slow: https://benhouston3d.com/blog/risc-v-in-2024-is-slow https://benhouston3d.com/blog/risc-v-in-2024-is-slow Once the speed issue is fixed (it needs to be roughly 10x faster than it currently is), then it will likely find a ton of applications, especially in SBC that are low power/low cost/good performance.
- Pet_Ant 2y ago
  AFAICT most ISAs themselves are not slow. Most modern ISAs can be implemented quickly because the decoder is not the bottleneck and high performance cores do so much. So it's more like a question of why hasn't someone with a high-performance core slapped a RISC-V ISA on it? Well, for one they weren't built with RISC-V in the first place so it would take some work to do and test. Second, there isn't currently a market for this. There is not consumer item that has RISC-V and could sell more units by getting a faster one. Third, a lot of cores are probably entangled with ARM IP for now. None of these are integral to RISC-V itself. If you could rewrite history and swap the ISAs for ARM and RISC-V it'd still have a headwind right now from the late start, and the tail wind of being license free. It's entirely non-technical.
  - bhouston 2y ago
    I completely agree that it is a relative "simple" matter to put a fast core in there, the techniques are known. You may not beat an Apple Mx chip, but you could at least get somewhere close to it, like 50% of the performance. And then RISC-V would take off like crazy.
    - Wytwwww 2y ago
      Who actually has fast core (>=50% of Apple M series) that they could just slap RISC-V on, though? ARM, Intel, AMD, Qualcomm? IIRC the Switch 2 chip is going to use Cortex so even Nvidia can't be bothered to design their own CPU cores. CPUs seem to be a commodity at this point so unless you're willing to invest massive amounts to be able to compete to catch up and overtake them just paying whatever ARM wants (probably pennies) seems like a good deal.
snvzz 2y ago
>Google had been making moves in this space – it officially supported RISC-V in the Android-specific, Linux-derived Android Common Kernel (ACK) up until May 2024 when it removed support from the ACK >*a significant setback for the architecture* Why do articles like this continue to repeat this baseless nonsense?

18 points by pseudolus 2y ago

Most people don't need more than a Chromebook has to offer. I can't imagine why this wouldn't eventually take off there and become the main CPU that people spend time on. Especially as margin tighten, the ISA-license will be will something that has to go. "Your margin is my opportunity".

Wytwwww 2y ago
IMHO it might be the complete opposite. Margins are low and and off the shelf ARM designs are very cheap so what's the point of making very risky investments when you can already get good enough off the shelf designs from ARM? If the licensing fee is only ~1-2% who can justify this besides Qualcomm, Apple, Nvidia, etc.? And Apple and Qualcomm only did this because they wanted to (and could afford to) design faster chips than ARM.
- snvzz 2y ago
  There's the option of licensing a RISC-V microarchitecture, rather than an ARM one. It has the advantage that RISC-V is an open standard, and there are many companies licenses for their own RISC-V microarchitectures, rather than being stuck with licensing ARM cores from ARM.
  - Wytwwww 2y ago
    From whom? I'm not sure why would anyone who can design high-end RISC-V cores would license them to someone else instead of making the chips themselves. As ARM has demonstrated that's not a very good business model (they are barely making any money compared to some of their customers).
    - snvzz 2y ago
      >From whom? From any of the IP vendors offering them. e.g. Tenstorrent has Ascalon (already licensed by LG) and Alastor, SiFive has P670 and P870, Akeana has the 5000 series.
      - Wytwwww 2y ago
        I just don't understand why would anyone who developed a competitive mid to high-end core would want to license it to anyone? What would they have to gain by doing that instead of making them themselves? Of course there are cases where the CPU is pretty much just a "cost centre" (embedded, industrial products and such) but I don't see that working in the PC/mobile/(other general purpose device) market. Doesn't really seem like a sustainable business model long-term i.e. you won't be able to generate enough revenue to compete with companies designing chips in house.
pjmlp 2y ago
Most people definitly need more RAM and SSD storage than most Chromebooks offer, specially when most middle range phones and tablets offer the double.
Someone 2y ago
For anything that sells for over $100 or so, I think not having to pay for the ISA license is nice, but the main benefit of RISC-V relative to ARM and the like is the flexibility of being allowed to add or remove features at will.
- Wytwwww 2y ago
  That's not relevant for anyone making Chromebooks (especially low to mid end ones) though. Unless you can get an equivalent RISC-V chip significantly cheaper than an off the shelf x86/ARM one why would they do it (considering the extra costs/overhead from supporting a niche platform in addition to lost sales).
whatevaa 2y ago
Hardware accelerated video decode is a must-have for any light chromebooks to gain traction.
- snvzz 2y ago
  There are many IP vendors to license video codec RTL from. RISC-V isn't a factor. Many RISC-V based SoCs already in the market have some level of hardware video codec acceleration built in.

bhouston 2y ago

I think that this article is missing the main reason why RISC-V is not yet mainstream. It is super slow: https://benhouston3d.com/blog/risc-v-in-2024-is-slow https://benhouston3d.com/blog/risc-v-in-2024-is-slow Once the speed issue is fixed (it needs to be roughly 10x faster than it currently is), then it will likely find a ton of applications, especially in SBC that are low power/low cost/good performance.

Pet_Ant 2y ago
AFAICT most ISAs themselves are not slow. Most modern ISAs can be implemented quickly because the decoder is not the bottleneck and high performance cores do so much. So it's more like a question of why hasn't someone with a high-performance core slapped a RISC-V ISA on it? Well, for one they weren't built with RISC-V in the first place so it would take some work to do and test. Second, there isn't currently a market for this. There is not consumer item that has RISC-V and could sell more units by getting a faster one. Third, a lot of cores are probably entangled with ARM IP for now. None of these are integral to RISC-V itself. If you could rewrite history and swap the ISAs for ARM and RISC-V it'd still have a headwind right now from the late start, and the tail wind of being license free. It's entirely non-technical.
- bhouston 2y ago
  I completely agree that it is a relative "simple" matter to put a fast core in there, the techniques are known. You may not beat an Apple Mx chip, but you could at least get somewhere close to it, like 50% of the performance. And then RISC-V would take off like crazy.
  - Wytwwww 2y ago
    Who actually has fast core (>=50% of Apple M series) that they could just slap RISC-V on, though? ARM, Intel, AMD, Qualcomm? IIRC the Switch 2 chip is going to use Cortex so even Nvidia can't be bothered to design their own CPU cores. CPUs seem to be a commodity at this point so unless you're willing to invest massive amounts to be able to compete to catch up and overtake them just paying whatever ARM wants (probably pennies) seems like a good deal.

snvzz 2y ago

>Google had been making moves in this space – it officially supported RISC-V in the Android-specific, Linux-derived Android Common Kernel (ACK) up until May 2024 when it removed support from the ACK >*a significant setback for the architecture* Why do articles like this continue to repeat this baseless nonsense?