5 ms·

What happened to the patent issues surrounding x86 emulation? For example I remember reading this: https://arstechnica.com/information-technology/2017/06/intel-

by simfoo 6y ago

What happened to the patent issues surrounding x86 emulation? For example I remember reading this: https://arstechnica.com/information-technology/2017/06/intel-fires-warning-shots-at-microsoft-claims-x86-emulation-is-a-patent-minefield/ https://arstechnica.com/information-technology/2017/06/intel... First Apple and now Microsoft. Do they license the patents from Intel or did they expire? Or are they working around that somehow?

mschuster91 6y ago
The original x86 patents are expired by now, the x86-64 patents are held by AMD.
- scoopertrooper 6y ago
  While that's true, Intel has steadily extended x86 over the years and many programs depend on those extensions, which are still patent protected.
  - Dylan16807 6y ago
    How many programs depend on SSE3 or later? Any earlier extensions are free now.
    - rasz 6y ago
      Games, pretty much all of the AAA ones. ~10 years ago you would read cries of AMD x64 users not able to play SSE3 games. 2-5 years ago situation was repeating with AMD Phenom I/II owners and SSSE3/SSE4.
      - Joe_Cool 6y ago
        Using a compiler with proper switches is hard it seems. Some also blame 3rd party libraries or DRM.
  - kalleboo 6y ago
    Rosetta 2 doesn't support the SIMD extensions AVX (2011) and newer, but compatibility still seems pretty good
    - andy_ppp 6y ago
      This is because very little uses AVX instructions, hasn't it always been better to optimise software for the GPU rather than AVX instructions, even though they are not perfectly interchangeable.
      - saagarjha 6y ago
        I think it's more that compatible Mac software has historically compiled for non-Haswell targets in some form or the other for compatibility.
      - acdha 6y ago
        You might be thinking of the most recent AVX-512 extensions which had significant power management complications delaying adoption. Many common programs have optimized SIMD paths because GPU computing is significantly more complicated and the setup overhead makes it slower for uses which don’t process tons of data - one of the interesting points about the M1 is that the unified memory design avoids the overhead of copying data between the CPU and GPU.
        andy_ppp 6y ago
        Yes sorry so am thinking of AVX-512 you are correct.
        acdha 6y ago
        No problem - Intel’s marketing hasn’t helped at all. I notice this mostly for cryptography- we do a lot of data validation so spending up SHA is a nice win.
    - rsynnott 6y ago
      Very little software would _depend_ on AVX being there; it'd be too great a limit on backwards compatibility.
Rexxar 6y ago
Are Intel claims valid somewhere ? They are probably not in Europe where patents generally cannot be used to prevent interoperability.
- orra 6y ago
  > patents generally cannot be used to prevent interoperability I think you're thinking of copyright? Where reverse engineering is allowed for interoperability.
  - DaiPlusPlus 6y ago
    No, it's patent-law, not copyright law. Interestingly while the US has Fair Use (which is an open-ended legal defence), I don't think any countries in the EU (or "Western Europe" if Brexit happens) have anything beyond the closed and enumerated "Fair Dealing" defence).
    - 0xffff2 6y ago
      >or "Western Europe" if Brexit happens Wait a second, one of us is behind the times. I thought Brexit happened earlier this year?
      - cogman10 6y ago
        Brexit is in a constant state of happening and not happening. You can think of it like a qubit.
        0xffff2 6y ago
        That was true for a long time, but to quote the Government of the Netherlands (the closest thing I can find to an official source) [0]: >The United Kingdom (UK) left the European Union (EU) on 31 January 2020. A transition period is now in place until 31 December 2020. As far as I can tell from my distant vantage point on the west coast of North America, there is very little doubt left. Brexit has already happened and most of the fallout will be settled by the end of this month. I supposed you could say that it's still in a constant state of happening, but it's well past the point of "not happening" as far as I can tell. [0]: https://www.government.nl/topics/brexit/brexit-where-do-we-stand https://www.government.nl/topics/brexit/brexit-where-do-we-s...
cesarb 6y ago
According to Wikipedia (https://en.wikipedia.org/wiki/X86-64#History https://en.wikipedia.org/wiki/X86-64#History), AMD's x86-64 was announced in 1999 and the full specification was released in 2000. Assuming patents last for 20 years, they should have all expired by now. For SIMD, this includes everything up to SSE2; most software which uses any of the newer SIMD extensions should have a fallback to SSE2, since older CPUs don't have these newer extensions.
- ksec 6y ago
  AH! Thank You. The timeline make sense. Seems Apple have time this perfectly for their transition. I am now giving them some creditable excuses for the negligence of Mac in the past few years.
  - sonotmyname 6y ago
    > Seems Apple have time this perfectly for their transition Important to note that this doesn't really apply to Apple Silicon. Rosetta II isn't X86 emulation, it's a translation layer - big technical difference.
    - tambourine_man 6y ago
      I get the technical difference but not the legal one. Why would one be OK but not the other?
      - temac 6y ago
        Especially since that kind of legal aspect highly depend on how much companies are willing to fight, and the state of mind of the judges. And in a not so distant past, Intel warned MS and PC vendors doing some ARM stuff that it might be willing to fight. As for the state of mind of the judges, it is difficult to predict... But it can be different. Depends on the magic formulas you use for your claims, and whether you are actually permitted to patent software in the first place (with or without other magic formulas usually stating that said software is executed by some hardware). And in places where it is not supposed to be possible to patent software, patent offices are usually nevertheless accepting them, because it gives them money so why not? :P TLDR it is a mess. Depends on an insane nmuber of contextual factors, and as unpredictable as the outcome of Oracle vs. Google before the verdict. Or maybe even more unpredictable.
iSnow 6y ago
Cross patent volume license deals are pretty much standard among the tech giants.
tsbinz 6y ago
MS had x86 emulation for a while, but not x64 (Windows on ARM could run 32-bit binaries, but not 64-bit).

by simfoo 6y ago

mschuster91 6y ago

The original x86 patents are expired by now, the x86-64 patents are held by AMD.

scoopertrooper 6y ago
While that's true, Intel has steadily extended x86 over the years and many programs depend on those extensions, which are still patent protected.
- Dylan16807 6y ago
  How many programs depend on SSE3 or later? Any earlier extensions are free now.
  - rasz 6y ago
    Games, pretty much all of the AAA ones. ~10 years ago you would read cries of AMD x64 users not able to play SSE3 games. 2-5 years ago situation was repeating with AMD Phenom I/II owners and SSSE3/SSE4.
    - Joe_Cool 6y ago
      Using a compiler with proper switches is hard it seems. Some also blame 3rd party libraries or DRM.
- kalleboo 6y ago
  Rosetta 2 doesn't support the SIMD extensions AVX (2011) and newer, but compatibility still seems pretty good
  - andy_ppp 6y ago
    This is because very little uses AVX instructions, hasn't it always been better to optimise software for the GPU rather than AVX instructions, even though they are not perfectly interchangeable.
    - saagarjha 6y ago
      I think it's more that compatible Mac software has historically compiled for non-Haswell targets in some form or the other for compatibility.
    - acdha 6y ago
      You might be thinking of the most recent AVX-512 extensions which had significant power management complications delaying adoption. Many common programs have optimized SIMD paths because GPU computing is significantly more complicated and the setup overhead makes it slower for uses which don’t process tons of data - one of the interesting points about the M1 is that the unified memory design avoids the overhead of copying data between the CPU and GPU.
      - andy_ppp 6y ago
        Yes sorry so am thinking of AVX-512 you are correct.
        acdha 6y ago
        No problem - Intel’s marketing hasn’t helped at all. I notice this mostly for cryptography- we do a lot of data validation so spending up SHA is a nice win.
  - rsynnott 6y ago
    Very little software would _depend_ on AVX being there; it'd be too great a limit on backwards compatibility.

Rexxar 6y ago

Are Intel claims valid somewhere ? They are probably not in Europe where patents generally cannot be used to prevent interoperability.

orra 6y ago
> patents generally cannot be used to prevent interoperability I think you're thinking of copyright? Where reverse engineering is allowed for interoperability.
- DaiPlusPlus 6y ago
  No, it's patent-law, not copyright law. Interestingly while the US has Fair Use (which is an open-ended legal defence), I don't think any countries in the EU (or "Western Europe" if Brexit happens) have anything beyond the closed and enumerated "Fair Dealing" defence).
  - 0xffff2 6y ago
    >or "Western Europe" if Brexit happens Wait a second, one of us is behind the times. I thought Brexit happened earlier this year?
    - cogman10 6y ago
      Brexit is in a constant state of happening and not happening. You can think of it like a qubit.
      - 0xffff2 6y ago
        That was true for a long time, but to quote the Government of the Netherlands (the closest thing I can find to an official source) [0]: >The United Kingdom (UK) left the European Union (EU) on 31 January 2020. A transition period is now in place until 31 December 2020. As far as I can tell from my distant vantage point on the west coast of North America, there is very little doubt left. Brexit has already happened and most of the fallout will be settled by the end of this month. I supposed you could say that it's still in a constant state of happening, but it's well past the point of "not happening" as far as I can tell. [0]: https://www.government.nl/topics/brexit/brexit-where-do-we-stand https://www.government.nl/topics/brexit/brexit-where-do-we-s...

cesarb 6y ago

According to Wikipedia (https://en.wikipedia.org/wiki/X86-64#History https://en.wikipedia.org/wiki/X86-64#History), AMD's x86-64 was announced in 1999 and the full specification was released in 2000. Assuming patents last for 20 years, they should have all expired by now. For SIMD, this includes everything up to SSE2; most software which uses any of the newer SIMD extensions should have a fallback to SSE2, since older CPUs don't have these newer extensions.

ksec 6y ago
AH! Thank You. The timeline make sense. Seems Apple have time this perfectly for their transition. I am now giving them some creditable excuses for the negligence of Mac in the past few years.
- sonotmyname 6y ago
  > Seems Apple have time this perfectly for their transition Important to note that this doesn't really apply to Apple Silicon. Rosetta II isn't X86 emulation, it's a translation layer - big technical difference.
  - tambourine_man 6y ago
    I get the technical difference but not the legal one. Why would one be OK but not the other?
    - temac 6y ago
      Especially since that kind of legal aspect highly depend on how much companies are willing to fight, and the state of mind of the judges. And in a not so distant past, Intel warned MS and PC vendors doing some ARM stuff that it might be willing to fight. As for the state of mind of the judges, it is difficult to predict... But it can be different. Depends on the magic formulas you use for your claims, and whether you are actually permitted to patent software in the first place (with or without other magic formulas usually stating that said software is executed by some hardware). And in places where it is not supposed to be possible to patent software, patent offices are usually nevertheless accepting them, because it gives them money so why not? :P TLDR it is a mess. Depends on an insane nmuber of contextual factors, and as unpredictable as the outcome of Oracle vs. Google before the verdict. Or maybe even more unpredictable.

iSnow 6y ago

Cross patent volume license deals are pretty much standard among the tech giants.

tsbinz 6y ago

MS had x86 emulation for a while, but not x64 (Windows on ARM could run 32-bit binaries, but not 64-bit).