Building a More Secure World with the RISC-V ISA

Re:Building a More Secure World with the RISC-V ISA by snvzz on Sunday January 07, 2018 @07:22AM (#16090602)

>Nowadays it just doesn't matter, because CPUs translate ISA to whatever they please.

Doing that comes at a cost, which compounds the complexity already added by merit of being CISC in the first place.

[ Parent | on HN ]

Re:Building a More Secure World with the RISC-V ISA by vardump on Sunday January 07, 2018 @08:01AM (#16090723)

Larger code size is not for free either.

Besides, this translation layer is not a major source of complexity anymore. Even instruction decoding in x86 is not a big deal anymore.

[ Parent | on HN ]

Re:Building a More Secure World with the RISC-V ISA by snvzz on Sunday January 07, 2018 @09:41AM (#16090980)

>Larger code size is not for free either.

What larger code size?

RISC-V is no worse than x86-64.

[ Parent | on HN ]

Re:Building a More Secure World with the RISC-V ISA by vardump on Sunday January 07, 2018 @10:16AM (#16091096)

That's a nice property for sure. Typically RISC ISAs have had larger code size.

Being a new ISA, RISC-V does have an advantage of being able to learn from past mistakes.

[ Parent | on HN ]

Re:Building a More Secure World with the RISC-V ISA by snvzz on Sunday January 07, 2018 @10:20AM (#16091108)

Simple vs Complex, but same code size. An easy win for simple.

It's very nice, indeed. Unlike the established RISC ISAs, this one's created with decades of RISC (and CISC) experience to draw from.

I'd of course still have my doubts whether this is going to draw any traction, if it wasn't for the list of RISC-V Foundation members.

[ Parent | on HN ]

Re:Building a More Secure World with the RISC-V ISA by microcolonel on Sunday January 07, 2018 @03:02PM (#16092365)

I think you somewhat underestimate the complexity introduced by decoding x86, and working with the resulting μ-ops.

[ Parent | on HN ]