Intel Discloses New CPU Instructions, Enhanced Hardware Feedback Interface (EHFI)

[gilboa]

You don't 'misunderstand' anything. Even the x86 experts here don't understand enough in order to answer as simple a question as that posed / posted in @#7.

Childish trolling aside, the full Intel developer manual can be downloaded from this URL:
https://www.intel.com/content/dam/ww...ual-325383.pdf

If you prefer the AMD version, feel free to download it here:

https://www.amd.com/system/files/TechDocs/24592.pdf

Free free to count the instructions yourself (as if the number of instructions means anything in this day and age...).

BTW, as someone that still write ASM code from time, I use both quite extensively. (I prefer the Intel variant) and both, are literally light years ahead of the horrible documentation available on ARM.

Oh, and by the way--if Intel has not seen fit to implement an interrupt system--starting with the 8080, and even before that--as sophisticated as that found in the Z80 and MC6800, you certainly don't think they're going to do it now, do you? The sheer burden of dealing with the complexities of having to save the environment---and of restoring it---of a "modern" x86 (everything is relative; some might say that "modern x86" is an oxymoron) makes that idea laughable, at the very least.

Granted, English is not my first language. But, umm, anyone care to translate the above?

- Gilboa

[jayN]

is that AMX or avx512 VNNI? Anyway, saturation is nice to have...

VPDPBUSDS

https://en.wikichip.org/wiki/x86/avx512_vnni

[danmcgrew]

User interrupts? Does that mean we get interrupts in user space like in micro controllers? Or do I misunderstand that completely?

You don't 'misunderstand' anything. Even the x86 experts here don't understand enough in order to answer as simple a question as that posed / posted in @#7.

***********************

Oh, and by the way--if Intel has not seen fit to implement an interrupt system--starting with the 8080, and even before that--as sophisticated as that found in the Z80 and MC6800, you certainly don't think they're going to do it now, do you? The sheer burden of dealing with the complexities of having to save the environment---and of restoring it---of a "modern" x86 (everything is relative; some might say that "modern x86" is an oxymoron) makes that idea laughable, at the very least.

[Alex/AT]

Adding some obscure stuff when even the current command set is in dire need of reducing complexity... well, wonder if it is a good idea.

[Lycanthropist]

User interrupts? Does that mean we get interrupts in user space like in micro controllers? Or do I misunderstand that completely?

[rastersoft]

Some time ago I read that AMD was suing Intel to ensure exclusive access to one of the three free "big blocks" of instructions that remains in the x86 architecture, in order to be able to add new instructions without risking clashing with the ones added by Intel. Does anybody know if they achieved it?

[bug77]

I'd like a definite answer, from someone who really knows---

How many instructions are now in the x86 instruction set?

All of them.

[tildearrow]

Phoronix Forums -- Now Sponsored by Intel

.....

However, we are not compensated to provide favorable opinions on products nor do we accept paid endorsements.

[danmcgrew]

I'd like a definite answer, from someone who really knows---

How many instructions are now in the x86 instruction set?

[jrch2k8]

Mmmm, EHFI sounds like exploiter's candy.

Let's hope Intel learned something this past couple of years