Intel Publishes "X86-S" Specification For 64-bit Only Architecture

[Anux]

You are all over the place. Please decide and make clear if you are talking about a hypotetical 64 bit only arch or the x86s from the article.

Leaving a few mode switches away at system boot does not increase efficiency, you are just falling for stupid marketing bullshit. It might reduce boot times by a few milliseconds and potentially make it easier for hard- and software devs in the future. Your link doesn't provide any benefit for the user.

[MorrisS.]

You are all over the place. Please decide and make clear if you are talking about a hypotetical 64 bit only arch or the x86s from the article.

Leaving a few mode switches away at system boot does not increase efficiency, you are just falling for stupid marketing bullshit. It might reduce boot times by a few milliseconds and potentially make it easier for hard- and software devs in the future. Your link doesn't provide any benefit for the user.

Your arguments are stupid. It's not marketing. A system with a simplified architecture is more efficient, moreover, a 64bit system is more safe from 32bit threats, and eventually 32bit intel CPUs will not disappear. I currently use AMD CPUs... I'm not an intel stupid fanboy, nontheless, I prefer pure 64bit cpus for more efficient purposes as well as operating systems should be only 64bit. 32 bit software has to die and company which uses 32bit software must fail.

[qarium]

SSE is a mix of 64-bit scalar instructions and 128-bit vector instructions. It is not a 128-bit instruction set, and you cannot implement a CPU using only them.
Announcing a processor with zero backward compatibility would be a disaster for AMD, since it would only gain adoption in niche applications like the itanium processor. It is very unlikely that AMD will try to repeat Intel’s mistake. It is also very unlikely that anyone will make a 128-bit processor anytime soon since 64-bit is good enough for general purpose computing. The main reason we went to 64-bit was for larger address spaces, and it will be decades before we run into the limitations of a 64-bit address space.
General purpose computing will likely never implement 1024-bit scalar support. There are only 2^166 atoms on earth and the upper estimate for protons/atoms in the universe is 2^272. No matter how much technology advances, the impetus behind 64-bit computing will never manifest for 1024-bit computing. Furthermore, the future shift to 128-bit computing is likely to be the last one we make. We might support bigger vector widths or go to some other computing model, but we will never implement higher scalar widths outside of niche applications.
Getting to the point where 128-bit address space is not enough would imply that we have constructed computer memory that is more than 5e11 kilograms in mass, and that assumes every bit is made out of a hydrogen atom and does not consider how we would read or write to them, or keep them in place. To illustrate just how absurd that is, 2^128 hydrogen atoms would exceed the mass of every navy in the world combined, by more than a factor of 10 million. The space needed would make a US aircraft carrier look like a grain of sand. It is just not realistic to think we will ever have a need for more than 128-bit hardware address spaces.

a 128bit or 256bit ISA/architecture would eliminate the use of PAE

https://en.wikipedia.org/wiki/Physical_Address_Extension

because large 64bit systems already use PAE to make larger ram address mapping possible.

right now you have one adress space ISA for cpus and one for GPUs and other usercases

128bit or 256bit ISA/architecture would make it possible to map every single important ISA command of the CPU and GPU architecture in one single 128bit/256 bit adress space. means no more 2 different ISAs for CPU and GPU...

the LibreSOC project did already hit this problem that 64bit was not enough to map all the needed commands because of this they developed simple-V

https://libre-soc.org/simple_v_extension/

an 128bit ISA/Architectur would make it possible to do this without the need of any extension.

[ryao]

a 128bit or 256bit ISA/architecture would eliminate the use of PAE

https://en.wikipedia.org/wiki/Physical_Address_Extension

because large 64bit systems already use PAE to make larger ram address mapping possible.

right now you have one adress space ISA for cpus and one for GPUs and other usercases

128bit or 256bit ISA/architecture would make it possible to map every single important ISA command of the CPU and GPU architecture in one single 128bit/256 bit adress space. means no more 2 different ISAs for CPU and GPU...

the LibreSOC project did already hit this problem that 64bit was not enough to map all the needed commands because of this they developed simple-V

https://libre-soc.org/simple_v_extension/

an 128bit ISA/Architectur would make it possible to do this without the need of any extension.

64-bit machines do not use PAE.

[qarium]

I've looked into FPGA options, you're right, an FPGA should be the solution, but the problem at the moment is the FPGA chips themselves. There is a 486 core, but it hasn't got a Floating Point Module. Anything newer than a 486 isn't currently supported on an FPGA due to the core size. The Current affordable FPGA chips aren't capable of running a Pentium CPU core. The ideal retro FPGA setup would be a Pentium2/Pentium3 CPU core with a Voodoo2 (or later Voodoo) and at least a SoundBlaster 16. If you're interested in the concept, checkout the AO486 CPU Core. It's been bundled into the MISTer Project.

well yes... but i am sure that we will see in the near future bigger FPGAs at a lower price.
but honestly you could only expect the complete minimum possible
similar to the 8bit world and 16bit world i believe there is a big market for retro 32bit gaming on FPGAs...

how big would a 486 be with a floating point unit and a voodoo2 and a soundblaster 16 ??

"If you're interested in the concept,"

of course i am... i think i will buy a mega65 and a VampirV4 to... and the 32bit world retro gameing is exactly my age and time.

i do not play modern games much. my first computer was a commodore 64 then a atari 1040st and then a macintosh performa 5200 and then a pentium 3 450mhz...

something like this: "486 be with a floating point unit and a voodoo2 and a soundblaster 16" would be very nice.

[qarium]

64-bit machines do not use PAE.

64 bit systems would need to use PAE as soon as we have more ram than 16EB of RAM...

or ???-...

i am pretty sure we will have a unified ISA/architecture with all the CPU+GPU space in one 128bit address pool...

"right now you have one adress space ISA for cpus and one for GPUs and other usercases
128bit or 256bit ISA/architecture would make it possible to map every single important ISA command of the CPU and GPU architecture in one single 128bit/256 bit adress space. means no more 2 different ISAs for CPU and GPU...
the LibreSOC project did already hit this problem that 64bit was not enough to map all the needed commands because of this they developed simple-V"

Simple-V is only needed because the GPU instcution ISA does not fit in the 64bit OpenPOWER ISA...

[ryao]

64 bit systems would need to use PAE as soon as we have more ram than 16EB of RAM...

or ???-...

i am pretty sure we will have a unified ISA/architecture with all the CPU+GPU space in one 128bit address pool...

"right now you have one adress space ISA for cpus and one for GPUs and other usercases
128bit or 256bit ISA/architecture would make it possible to map every single important ISA command of the CPU and GPU architecture in one single 128bit/256 bit adress space. means no more 2 different ISAs for CPU and GPU...
the LibreSOC project did already hit this problem that 64bit was not enough to map all the needed commands because of this they developed simple-V"

Simple-V is only needed because the GPU instcution ISA does not fit in the 64bit OpenPOWER ISA...

There are no single address space machines with 1EB of RAM, much less 16EB. We would be close to a 128-bit transition if people were building those. The most that I have ever seen on a spec sheet is 64TB of RAM:

https://www.hpcwire.com/2012/06/14/sgi_launches_second_generation_uv_supercomputer/

[qarium]

There are no single address space machines with 1EB of RAM, much less 16EB. We would be close to a 128-bit transition if people were building those. The most that I have ever seen on a spec sheet is 64TB of RAM:
https://www.hpcwire.com/2012/06/14/s...supercomputer/

right. but maybe future designs are not about "single machines" ?? as other alraedy said(not my idea) with a bigger address space you could map the complete internet and all the computers in the world into one single address space.

[ryao]

right. but maybe future designs are not about "single machines" ?? as other alraedy said(not my idea) with a bigger address space you could map the complete internet and all the computers in the world into one single address space.

That is such a niche idea that nobody is going to build it. PAE remains a 32-bit hack that has nothing to do with 64-bit.

[Weasel]

right. but maybe future designs are not about "single machines" ?? as other alraedy said(not my idea) with a bigger address space you could map the complete internet and all the computers in the world into one single address space.

Even on a single machine you map less on a single process since each process gets its own address space.

And you want to map the entire internet into one process.

Nice security.