5 ms·

> RISCV architecture (a free architecture that doesn't include that shit) Surely you can't think the architecture itself is the differentiator. x86 didn't have

by judge2020 5y ago

> RISCV architecture (a free architecture that doesn't include that shit) Surely you can't think the architecture itself is the differentiator. x86 didn't have all of this security 20 years ago, give engineers a few years of time to throw some locks on a risc-v chip and it'll be Enterprise Ready™ in no time.

orangepurple 5y ago
The good news is the main manufacturers of RISC-V are Chinese vendors that allow complete access to low level processor details. They generally don't lock down their products at all.
- pjmlp 5y ago
  Bad news is that US doesn't want to have anything with them.
  - orangepurple 5y ago
    You can still import them into the US en masse
    - pjmlp 5y ago
      Still, until the government plays an Huawei move.
    - avianes 5y ago
      Not if they break US patents (which is likely), or if the US puts more constraints on imports.
  - voakbasda 5y ago
    At this point, their government has completely undermined foreign confidence in their semiconductor industry. I would be careful to avoid any processors from there, because the chance that it contains a backdoor is just too great to risk. Even if that is only a perceived risk and not a real one. American propaganda in recent years clearly set out to reach this new status quo, and it’s probably all smoke and little to no fire. But it worked. There are enough examples of China screwing over foreign companies to prove the risk could be real no matter your size (e.g. ARM China).
    - pjmlp 5y ago
      That applies to every foreign country actually. Anyone that wants to be 100% sure of the supply chain has to move away from globalization.
      - hilbert42 5y ago
        Exactly! Exactly! Exactly! I still cannot understand why the US and other countries with high tech allowed themselves to become so absolutely dependent on China. When their own companies killed of local manufacturing and went to China out of greed and increased profits then governments should have stepped in on strategic grounds, so why didn't they? It was obvious to me some 25 years ago and I've no monopoly on this insight so they must have been well aware that this would happen. In essence, these countries have been shooting themselves in the feet for decades. The question is why. Edit: I can remember the time when the US military required certain components, chips etc. to be able to be second-sourced from multiple manufacturers before they were incorporated into equipment. Does anyone else remember this? Clearly governments, not only the US, have been aware of the problem for many decades and have chosen to do nothing about it. Moreover, what brought them to abandon this once good policy in the first instance?
    - throw58765434 5y ago
userbinator 5y ago
With the (already?) expiration of x86 patents, I'd love to see a "pure" x86 implementation without any of the user-hostile crap, and see how far the community can take it; but sadly, the RISC bandwagon is diverting attention away from that. A CPU without the user-hostile features but still able to run the massive existing software base would be ideal.
- alerighi 5y ago
  Would be too difficult to implement. x86 is a very big instruction that is impossible to implement with an hardware: both Intel and AMD processors in fact run inside a virtual machine that translates x86 instructions in an internal RISC instruction set that is manageable by the real CPU architecture. If Apple decided to move away from x86 and go to ARM to have their processor, and we are talking about one of the biggest companies, I don't think any community project will ever succeed in doing another x86 compatible CPU. On the other side RISCV instruction set is far simpler, being a RISC instruction set it decides to not have advanced optimizations in the processor (even better, none at all) and leave the optimization work to the compiler, that not only simplifies the processor, but also reduces the surface of attack of the processor (Meltown, Spectre, and all these attacks are just impossible on RISCV!). Of course that has a performance penalty, but since you simplify the processor you can just put more core in the saved space right?
  - userbinator 5y ago
    I'm not sure if you're being satirical, but open source x86 cores do exist --- they're around a 486 in terms of compatibility. Look up ao486 for example. What I'm referring to is the expiration of patents from the P6 era, which would mean all the uop-based stuff is now free to implement. What a lot of the RISC hype doesn't understand is the huge value in backwards compatibility --- you can have your "100% free" world but it'll forever remain niche. We need to accommodate the proprietary world if we want any chance of freedom winning; and not try to divide the world of computing.
    - coldtea 5y ago
alerighi 5y ago
RISCV is an open architecture. If a manufacturer does that, simply don't buy the processor from that manufacturer and buy it from another. All your software will still be compatible since it's the same architecture. Otherwise with x86 is more complex: you can choose between Intel and AMD (that has bought the license for the x86 instruction set - not something cheap to get), and both of them had their backdoor processor inside the computer (at least on Intel there are ways to disable it, as far as I know with AMD is more difficult if not impossible to do).
- pjmlp 5y ago
  Assuming that the software is all available from source and can be recompiled. Only the base RISC V is guaranteed thanks extensions. Also you are forgetting that just like Android and ARM, there are other forces at play that don't make it as easy in practice as FOSS advocates wish for.
  - smoldesu 5y ago
    > Assuming that the software is all available from source and can be recompiled. I remembered hearing that same line when I bought a Raspberry Pi in 2012. "It's useless! You can't run x86 software on it, so what's the point?" Flash-forwards a decade and now Graviton instances are blowing up like nothing else in the industry. RISC-V is in a very similar position to ARM 10 years ago; the groundwork has been laid, standards have been ratified and base packages/several kernels work perfectly fine on it. The only difference is that ARM is more expensive to license and is less flexible. > Only the base RISC V is guaranteed thanks extensions. Yeah. Is that a problem? The situation on ARM is equally bad if not worse (frequent iterations end up throwing even relatively recent CPU models under the bus), and the reason why RISC-V divided itself into extensions is so that you didn't have to start from scratch when John RISC decides to add in 3 new floating point instructions. It's a pretty damn good compromise if you ask me, and it certainly doesn't have any bearing on software availability; RISC-V programs run on RISC-V processors. ARM does not have that same liberty. There are plenty of genuine constraints for RISC-V (the majority of them in the manufacturing/mass production side of things, now), but the majority of these software issues have been solved and taped out years ago.

orangepurple 5y ago

The good news is the main manufacturers of RISC-V are Chinese vendors that allow complete access to low level processor details. They generally don't lock down their products at all.

pjmlp 5y ago
Bad news is that US doesn't want to have anything with them.
- orangepurple 5y ago
  You can still import them into the US en masse
  - pjmlp 5y ago
    Still, until the government plays an Huawei move.
  - avianes 5y ago
    Not if they break US patents (which is likely), or if the US puts more constraints on imports.
- voakbasda 5y ago
  At this point, their government has completely undermined foreign confidence in their semiconductor industry. I would be careful to avoid any processors from there, because the chance that it contains a backdoor is just too great to risk. Even if that is only a perceived risk and not a real one. American propaganda in recent years clearly set out to reach this new status quo, and it’s probably all smoke and little to no fire. But it worked. There are enough examples of China screwing over foreign companies to prove the risk could be real no matter your size (e.g. ARM China).
  - pjmlp 5y ago
    That applies to every foreign country actually. Anyone that wants to be 100% sure of the supply chain has to move away from globalization.
    - hilbert42 5y ago
      Exactly! Exactly! Exactly! I still cannot understand why the US and other countries with high tech allowed themselves to become so absolutely dependent on China. When their own companies killed of local manufacturing and went to China out of greed and increased profits then governments should have stepped in on strategic grounds, so why didn't they? It was obvious to me some 25 years ago and I've no monopoly on this insight so they must have been well aware that this would happen. In essence, these countries have been shooting themselves in the feet for decades. The question is why. Edit: I can remember the time when the US military required certain components, chips etc. to be able to be second-sourced from multiple manufacturers before they were incorporated into equipment. Does anyone else remember this? Clearly governments, not only the US, have been aware of the problem for many decades and have chosen to do nothing about it. Moreover, what brought them to abandon this once good policy in the first instance?
  - throw58765434 5y ago

userbinator 5y ago

With the (already?) expiration of x86 patents, I'd love to see a "pure" x86 implementation without any of the user-hostile crap, and see how far the community can take it; but sadly, the RISC bandwagon is diverting attention away from that. A CPU without the user-hostile features but still able to run the massive existing software base would be ideal.

alerighi 5y ago
Would be too difficult to implement. x86 is a very big instruction that is impossible to implement with an hardware: both Intel and AMD processors in fact run inside a virtual machine that translates x86 instructions in an internal RISC instruction set that is manageable by the real CPU architecture. If Apple decided to move away from x86 and go to ARM to have their processor, and we are talking about one of the biggest companies, I don't think any community project will ever succeed in doing another x86 compatible CPU. On the other side RISCV instruction set is far simpler, being a RISC instruction set it decides to not have advanced optimizations in the processor (even better, none at all) and leave the optimization work to the compiler, that not only simplifies the processor, but also reduces the surface of attack of the processor (Meltown, Spectre, and all these attacks are just impossible on RISCV!). Of course that has a performance penalty, but since you simplify the processor you can just put more core in the saved space right?
- userbinator 5y ago
  I'm not sure if you're being satirical, but open source x86 cores do exist --- they're around a 486 in terms of compatibility. Look up ao486 for example. What I'm referring to is the expiration of patents from the P6 era, which would mean all the uop-based stuff is now free to implement. What a lot of the RISC hype doesn't understand is the huge value in backwards compatibility --- you can have your "100% free" world but it'll forever remain niche. We need to accommodate the proprietary world if we want any chance of freedom winning; and not try to divide the world of computing.
  - coldtea 5y ago

alerighi 5y ago

RISCV is an open architecture. If a manufacturer does that, simply don't buy the processor from that manufacturer and buy it from another. All your software will still be compatible since it's the same architecture. Otherwise with x86 is more complex: you can choose between Intel and AMD (that has bought the license for the x86 instruction set - not something cheap to get), and both of them had their backdoor processor inside the computer (at least on Intel there are ways to disable it, as far as I know with AMD is more difficult if not impossible to do).

pjmlp 5y ago
Assuming that the software is all available from source and can be recompiled. Only the base RISC V is guaranteed thanks extensions. Also you are forgetting that just like Android and ARM, there are other forces at play that don't make it as easy in practice as FOSS advocates wish for.
- smoldesu 5y ago
  > Assuming that the software is all available from source and can be recompiled. I remembered hearing that same line when I bought a Raspberry Pi in 2012. "It's useless! You can't run x86 software on it, so what's the point?" Flash-forwards a decade and now Graviton instances are blowing up like nothing else in the industry. RISC-V is in a very similar position to ARM 10 years ago; the groundwork has been laid, standards have been ratified and base packages/several kernels work perfectly fine on it. The only difference is that ARM is more expensive to license and is less flexible. > Only the base RISC V is guaranteed thanks extensions. Yeah. Is that a problem? The situation on ARM is equally bad if not worse (frequent iterations end up throwing even relatively recent CPU models under the bus), and the reason why RISC-V divided itself into extensions is so that you didn't have to start from scratch when John RISC decides to add in 3 new floating point instructions. It's a pretty damn good compromise if you ask me, and it certainly doesn't have any bearing on software availability; RISC-V programs run on RISC-V processors. ARM does not have that same liberty. There are plenty of genuine constraints for RISC-V (the majority of them in the manufacturing/mass production side of things, now), but the majority of these software issues have been solved and taped out years ago.