Reply to post: Re: Some serious questions.

Will this be one of the world's first RISC-V laptops?

juice

Re: Some serious questions.

> However x86 and to a lesser extent also ARM carries some baggage due to their age. Furthermore they both contain everything you could dream of in a general purpose processor, and more

It's worth bearing in mind that RISC-V was introduced 12 years ago, and is based on David Patterson's RISC designs, which he first drew up in 1990 while in academia. So it's not *that* new, though at least RISC-V was effectively a "clean slate", since it didn't have to carry any significant baggage over from its school days.

> RISC-V is super simple and most functionality, even floating point operations, is optional extensions

You can't add or remove things from an x86 chip, but ARM very much lets you pick and choose what you want in your SoC.

In fact, that's part of the reason why Apple's ARM chips are dominating things at present, since they've picked the bits they want and then done their own engineering and custom design work atop that!

https://www.arm.com/products/custom-socs

https://www.arm.com/technologies/custom-instructions

It's also worth bearing in mind that all that extra flexibility carries costs of it's own. One reason why ARM took so long to make any inroads against x86 dominance of the "PC" market was that while there were millions (if not billions) of ARM devices out there, each one was based on a unique design and OS/software had to be custom tailored to each one.

It wasn't until we got Android (and to a degree, iOS) that we started to get properly standardised ARM hardware designs which you could build standardised software packages for. And that then led to things like the Raspberry Pi and it's ilk.

> It’s also open, so companies wouldn’t have to pay for a licence, in theory, because they would in most cases still have to buy a design somewhere… Maybe there would be companies offering custom designs to the industry.

It'll be interesting to see how pricing pans out over time; for most "commercial" Open Source stuff, the main charges come from support and training, and I'd be surprised if those costs were roughly on par with those for x86 and ARM chips.

> Also because it is so new and carry so little baggage, general purpose versions for PC’s have the potential to have very good power/performance. It should at least be able surpass x86 to rival ARM.

Perhaps. The last article I saw about such things indicated that while a prototype RISC chip was significantly beating the competition in the power stakes, it was also only running about a quarter of the speed.

And even then, that comparison was based on benchmark results provided by the manufacturer and using a deliberately simplified single-process benchmark, since Ars didn't actually have a sample of the RISC chip to test.

And we all know just how reliable and unbiased manufacturer-supplied benchmark figures can be!

https://arstechnica.com/gadgets/2020/12/new-risc-v-cpu-claims-recordbreaking-performance-per-watt/

Fundamentally, RISC-V may be improving faster than it's competition, but that's partly because it's so far behind them. And it remains to be seen whether it'll be able to become truly competitive on both price and performance, especially given that x86 and ARM both get a lot more design-money thrown at them, and are able to licence patents etc.

On the other hand, there's plenty of niches where low-power (and/or patent-free/politically unencumbered) CPUs can be slotted into. So I think RISC-V is here to stay, regardless!

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon