Live data from Hacker News

RISC-V AI Chips Will Be Everywhere (2022)

spectrum.ieee.org

1–10 of 21 posts

Re: RISC-V AI Chips Will Be Everywhere (2022)

#5
post #2

I don't know. It feels to me at this point like the RISC-V revolution and the year of Linux on the desktop will happen at around the same time.

I agree with you. If you read the original presentation by Dave Patterson that introduced RISC-V, the idea was to unify disparate architectures ( microcontrollers, DSPs, CPUs) to unify their software ecosystem (compilers, optimizers, debuggers) to facilitate and accelerate development.

I don't think that goal has yet been achieved. Instead, as others point out, we've seen a plethora of offerings from cheap vendors dodging the ARM "monopoly".

In the AI space, established vendors like Nvidia & AMD have no incentive to overhaul their entire stack to RISC-V. It's a huge investment and risk for unknown returns.

Disruption will have to come from outside, but it'll be a while before we see Nvidia dethroned.

Re: RISC-V AI Chips Will Be Everywhere (2022)

#8
post #4
post #2

I don't know. It feels to me at this point like the RISC-V revolution and the year of Linux on the desktop will happen at around the same time.

How long has RISC-V been around? Seems to me that it's doing well on any measure.

Are you counting "real-world usage" as a measure? I can go to Costco or Best Buy right now and come home with a system with an ARM or x86-64 chip. I'd have to go out of my way to find a system with RISC-V at its core. Even if you allow for devices like routers or embedded systems in disk drives and the like, there's a good chance the selection that any given store has won't have a RISC-V.

Maybe you can come back to this comment in five or ten years and laugh at me, but right now it feels like RISC-V is a solution in search of a problem.

Re: RISC-V AI Chips Will Be Everywhere (2022)

#9
post #2

I don't know. It feels to me at this point like the RISC-V revolution and the year of Linux on the desktop will happen at around the same time.

No chance. Linux on the desktop will never happen because of fundamental structural issues (lack of funding, no motivation for open source devs to make it good enough for normal people, many naysayers in the community not wanting it to work well for normal people).

RISC-V doesn't have any structural issues that would mean it can't succeed. In fact it already is very successful, but it's mostly in embedded and invisible CPUs at the moment which is why it might not seem like it.

It definitely needs some time to mature for application class processors but I can't see any reason why it won't.

Re: RISC-V AI Chips Will Be Everywhere (2022)

#10
Most people don't realize that most chips have more processors than just the main CPU cluster. The PCIE and DDR PHYs often have tiny low end CPU cores used for things like link training. There may be 10 of those CPU cores and these are the kinds of things that are being replaced by RISC-V. You don't need a fast CPU for this stuff.

Eventually RISC-V may have CPU cores fast enough to compete with high end ARM and x86 cores but it will take a while and incentive from the chip companies to design those high performance RISC-V cores. Some of that incentive may come from ARM raising license fees.

Post reply on HN