RISC-V: You Should Be Looking Into It Today

If you'd have told me five years ago that soon enough I will be able to drop a Cortex M IP in my design, at no price at all, there is a good chance that I would have said something snarky about optimism and how harsh reality actually is.

And yet here we are. Arm just started giving Cortex M IP cores away.

This move was not unexpected. Indeed, so far, the most compelling arguments to use something other than Arm's cores had to do with the the royalties and the legal complications behind them. Most companies that integrate IPs in programmable logic designs have some experience in dealing with the legal aspects, and there is only so much that a company which basically lives on licensing IP can do about the legalese. The price, then, was the most obvious thing to slash in order to gain an edge over the up-and-coming adversary.

What is unexpected is how quickly it happened. Granted, it's been almost an year since Western Digital announced that it was going to use RISC-V cores, but this is a pretty short timeframe for companies of these sizes, and designs of these complexities, spanning multiple product lines. One can only speculate about whether this is a proactive move or a reaction to an exodus of greater proportions than we realize.

Either way, when a company like Arm willing to release IP for free, even under fairly restrictive licensing agreements (most of which are fairly legitimate, to be fair), it's a sign that the competition is real.

RISC-V isn't a theoretical risk anymore, something that customers are just investigating, and which you can "address" with a smear site. When the management of a company like Arm decides to let go of some revenue, you can bet there were quite a few customers who not only investigated RISC-V, but also liked what they found.

It means that there are important commercial vendors for whom RISC-V is viable today. RISC-V is past the stage where it's only for risk-taking innovators.

Which means that, if you have been planning to play with a RISC-V design lately, but just didn't find the time for it, it's probably a good idea to find the time, and fast.