Hacker News Reader

Comment by pier25 | Hacker News Reader

Risc-v is already coming for the data center and Apple have their hands full.

RISC-V is at-best 10-15 years away from being in the position to where ARM is now.

That isn't how CPU design works. The microarchitecture is what matters and you can use the same techniques as x86 uses today to get good performance.

Just because there is a new ISA does not mean a company cannot design a modern CPU.

For example, the Vision Star Five is uses an in order CPU core that is competitive with the out of order A72 core. The performance gap on day one is much smaller. It also has UEFI support and opensource GPU drivers for Power VR are in development by imagination, so the software support will also quickly catch up or surpass ARM SBCs.

The same will happen in the server market. Nobody who works on a RISCV server chip has to wait for ARM to deliver a CPU core that doesn't suck.

Likewise, because there is a modern CPU, doesn't mean the relevant partners that actually matter in the industry care enough about its existence.

Or that they show up in computer shops for regular buyers, instead of computer enthusiasts, used to install GNU/Linux into any kind of motherboard with a CPU on it.

I have no insights about this but Jim Keller disagrees with that statement

Jim Keller is someone to pay attention to, but as a sibling says, JK has a commercial IP interest, which means the touted openness advantage could easily be lost. This insight comes from people who have consulted for organisations like Tenstorrent and SambaNova and both companies focussed on designing for HPC and AI. General-purpose consumer computing platform (i.e. servers, desktops and laptops) is yet to be forthcoming from anyone in the space, and they are a way off.

> which means the touted openness advantage could easily be lost

RISC-V openness has nothing to do with processor implementation, it's solely in the instruction set.

The advantage it has is that it provides compatibility across vendors and anyone can become a vendor of RISC-V IP. Comercial IP doesn't affect that, and if anything makes multiple IP vendors more likely.

The availability of server processors depends on the market rather than any technical limitation. If someone decides to take the risk of investing in the development of a strong server product then we'll get it. It depends on some company believing there is a market for it.

> which means the touted openness advantage could easily be lost

I don't expect them to open source their processors, but they do definitely contribute back to the community. E.g. they open sourced their first risc-v vector implementation (ocelot) on top of BOOM. It's in order and they used it as a reference for their later out of order commercial designs, but they also could've keep it to them self's.

JK got vested interested in such a wishful thinking statement. I take it as marketing and not engineering.

You can have a vested interest in something and still say the truth. Its not mutually exclusive.

Couldn't it be the other way around? That JK got into RISC-V because of its promising future?

In the overall market, probably, but in the data center?

Yes, and soon we will have PIC powered and AVR powered servers in the data center, joining our ARM and RISC V servers.

Don't forget the 6502.xxlarge instance coming soon to AWS :) /s

I'm betting the shop on 68k.24xlarge

You can have whatever you want, via FPGA instances (like F1).

There are ARM and RISC-V servers. The ARM ones are even available in major clouds. Why are you comparing to microcontrollers?

> RISC-V servers

There are?

I'm counting the HiFive; I suppose it depends how you define your terms. The point is that RISC-V can scale down to compete with PIC/AVR, but it can also scale up to "real" compute so it's a bit nonsense to dismiss it as only being for embedded.

> I'm counting the HiFive

Isn't it years behind ARM/x86 performance wise?

> nonsense to dismiss it as only being for embedded

They unlikely to be competitive anywhere else without significant investment.

> but it can also scale up to "real" compute

Did anyone doubt that you could theoretically scale up ARM to be competitive with x86 back in ~2000? It still took them several decades to achieve that..

> but it can also scale up to "real" compute so it's a bit nonsense to dismiss it as only being for embedded.

Especially after being selected for EU super computer effort.

I'm still betting on Intel bringing back Itanium.

Cause why not.. having more architectures (for no apparent reasons) sounds cool.

Keep dreaming.

Reply on news.ycombinator.com