Qualcomm, One of Arm's Biggest Customers, Starts a RISC-V Joint Venture

An anonymous reader quotes a report from Ars Technica: Arm is facing down its biggest competition ever, with the up-and-coming RISC-V architecture threatening to unseat it as the CPU at the center of almost every portable device. Now, one of Arm's biggest customers is trying out RISC-V, as Qualcomm is getting involved in a joint venture dedicated to the architecture. The joint venture doesn't have a name yet, but Qualcomm, NXP, Nordic Semiconductor, Bosch, and memory giant Infineon are all teaming up to form a new company that Qualcomm's press release says is "aimed at advancing the adoption of RISC-V globally by enabling next-generation hardware development." At first, the group will be focused on automotive uses, with an "eventual expansion" to IoT and mobile, Qualcomm's biggest market. Further reading: Qualcomm Chip Sales Down 25 Percent, Plans Layoffs

ARM is making this easy

[thatseattleguy]

ARM is trying to pump up its revenue in advance of its IPO later this year and someone in their Csuite had the brilliant idea of suing Qualcomm (regarding its acquisition of Nuvia) because...profit?
.

But maybe not so much. Turns out suing your biggest customers makes them look around even harder for alternatives to your product - especially an open-source, royalty-free one like RISC-V. So ARM's moves here aren't looking so much "brilliant" - more like "play stupid games, win stupid prizes".

That's in addition to RISC-V being a more modern and arguable more efficient and extensible architecture, in addition to its clear legal advantages. So while the writing is on the wall, ARM just put more bullets in its own foot-gun.

Why would Apple want ARM?

[drnb]

No big deal. Someday, Apple will just buy ARM outright and then Apple will continue having their happy little proprietary walled garden all to themselves.

Apple was one of the members in the founding consortium of ARM. They have a permanent license to anything ARM every produces.

What would buying ARM do for them? If they wanted to sell Apple Silicon chips they could already to that.

Re:

[nicubunu]

What would buying ARM do for them?

Ensure ARM stays alive after many of their customers move to RISC-V

Re:

[drnb]

What would buying ARM do for them?

Ensure ARM stays alive after many of their customers move to RISC-V

Why? Apple has access to all existing ARM IP and always will. Apple designs advanced ARM variants independent of ARM. ARM lags Apple.

Re:

[laughingskeptic]

ARM is where it is because of its efficiency. Both architectures are equally modern, have to give the extensibility edge to the open source architecture though ARM has been extended by all their high-end customers. The problem with RISC-V is it is a little too flexible in its spec which means libraries will generally not be reusable and if too much advantage is taken of the flexibility you are going to have a never ending pile of unique bugs.

Re:

[tlhIngan]

The problem with RISC-V is it is a little too flexible in its spec which means libraries will generally not be reusable and if too much advantage is taken of the flexibility you are going to have a never ending pile of unique bugs.

This is actually way bigger than you might think.

To call yourself a RISC-V processor, all you have to implement is the integer 32-bit core. That's it. RISC-V you are. That's why there are plenty of TTL logic chip implementations of RISC-V - when all you need is a basic core instr

Re:

[rickb928]

Oh, so like ARM, where Raspberry Pi binaries might have trouble with Snapdragon binaries, or or a Snapdragon 8cx having trouble say a 400...

As if ARM is a monolith. I, for one, welcome our flexible RISC-V overlords. May the distros divide and conquer!

Re:

[Misagon]

Yes, but I expect Qualcomm's goal is to produce chips for running Android.
Android for RISC-V has already an extensive feature requirement list. For instance, it requires vector and vector cryptography extensions.

There's also an independent standardisation effort for creating "Applicaton Profiles" for what a RISC-V "application-class" processor should support. Those are quite aligned with Android's requirements already. These also standardise certain behaviours, such as allowing misaligned loads and requirin

Spec flexibility

[Alwin Henseler]

I've been following RISC-V developments for a while now, and what's happening gives reason for optimism in this regard. Some observations:

Vector extension(s?) are a hot topic among implementers. Not surprising, given the big impact they can have on common audio/video codecs. There are some pre-spec implementations out there, but compiler support for those is (and likely, remains) poor. So it's safe to expect implementations out in the field 5..10 years from now, will be ones following post-frozen spec, w

Why wouldn't they go with RISC

[wakeboarder]

When your design company can be bought out and you can't use it anymore? I think everyone is rethinking ARM a little bit more these days.

From yesterday.

[waspleg]

Qualcomm -25% sales YoY, -52% net income YoY, already cut jobs, says more cuts coming, lowered next quarter guidance. [arstechnica.com]

Re:

[holostagram]

Sounds like a company that has committed to making some changes.