News: 1671103571

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Qualcomm talks up RISC-V, roasts 'legacy architecture' amid war with Arm

(2022/12/15)


Comment As Qualcomm tries to fight off a lawsuit from Arm demanding Qualcomm destroy its custom cores, the Snapdragon giant has signaled it may have a bigger future with RISC-V.

And that's all while Qualcomm criticized an "existing legacy architecture" for having useless features and not meeting certain requirements.

At the [1]RISC-V Summit this week, Qualcomm Director of Product Management Manju Varma said RISC-V, an emerging alternative to the proprietary Arm instruction set architecture, has opportunities across a range of devices Qualcomm designs chips for, from wearables and smartphones to laptops and connected cars.

Qualcomm pushes latest Arm-powered Snapdragon chip amid bitter license fight [2]READ MORE

While Qualcomm continues to use Arm's instruction set architecture (ISA) and off-the-shelf CPU designs as the basis for the application processing cores inside its system-on-chips (SoC), the US giant turned to RISC-V for microcontroller cores within its chips, starting with the [3]Snapdragon 865 SoC in 2019, according to [4]Varma , who helps drive Qualcomm's CPU strategy and roadmap across the company's entire portfolio.

Varma said the Snapdragon giant now uses RISC-V microcontrollers in SoCs in PCs, mobile devices, wearables, connected cars, as well as augmented reality and virtual reality headsets. These microcontrollers perform low-level work in the background, such as managing hardware.

[5]

This has resulted in Qualcomm apparently shipping more than 650 million RISC-V cores to date, making the ISA "one of the core technologies for Qualcomm" and Qualcomm "one of the leaders of RISC-V implementation."

[6]

[7]

Qualcomm is, we note, a founding member of RISC-V International, the body that oversees the royalty-free, open source ISA, so its support and enthusiasm for RISC-V is not a surprise.

The Qualcomm exec added that RISC-V has opportunities for use cases with higher value than microcontrollers.

[8]

"Now having a common base instruction set architecture that can scale from low-end microcontrollers, all the way to high-performance compute, and anything in between, it really enables efficiencies in the overall industry," Varma said in her keynote.

The claimed benefits of RISC-V over the 'legacy architecture'

Qualcomm first turned to RISC-V for a microcontroller in the Snapdragon 865 because it "needed something that was customizable, met our unique requirements, and had a small footprint," according to Varma. One thing about RISC-V is that it can be extended with custom instructions and features by CPU core implementers.

"The solutions from the existing legacy architecture didn't meet these requirements," she added.

Varma said the major benefit of RISC-V is that it receives feature contributions from various companies and organizations in all layers of the "value chain," from the ISA and CPU, to the system software, to the operating system, to the end-user applications. This contrasts with the legacy architecture, which is "owned by one entity in the value chain," she added.

The contribution structure for RISC-V, enabled by its open-source nature and governed by the nonprofit RISC-V International, creates an "opportunity to add features that add value to end consumers [and] are defined in concert with everyone in this value chain," according to Varma.

[9]

This is another area where the legacy architecture has fallen short, she added.

"Now, frequently we've seen in the past, with the legacy architecture, where features were introduced that didn't really seem to add value to end consumers," Varma said.

With RISC-V, there's an opportunity to define chip designs that have "best-in-class performance, best-in-class power efficiency, and value-added features," she said.

So wait, is Arm the 'legacy architecture?'

In the three times Varma referenced this "legacy architecture," she didn't call it by name. While we can't say definitively that she is talking about Arm, there are few good reasons to believe that she is:

Qualcomm has historically been a major licensee of Arm's ISA and off-the-shelf chip designs.

Arm has claimed that it's a [10]dominant player in the microcontroller space, with its Cortex-M designs accounting for [11]nearly three quarters of Arm-based chip shipments every year.

Varma compared RISC-V to the legacy architecture in the context of an ISA that can scale from microcontrollers to high-performance computing, which represents the range of Arm's capabilities. As a result, we don't think she was talking about the many other ISAs out there for microcontrollers. And we didn't get the impression she was talking about x86.

Then there's the fact that [12]Arm is suing Qualcomm in an attempt to destroy the latter's custom Nuvia cores, which are designed to be compatible with the former's ISA.

Arm believes Qualcomm's custom cores should be destroyed because, in Arm's view, the Snapdragon giant failed to negotiate a new architectural license agreement after it [13]acquired the startup Nuvia , where development of the cores began, in 2021.

Qualcomm, on the other hand, [14]argues it can continue developing the custom cores because it has existing architectural license with Arm that is "broadly overlapping" with the one Nuvia had.

While Qualcomm hasn't said what it will do if it loses the lawsuit, [15]multiple [16]analysts have said the legal feud is giving companies more reason to consider RISC-V as an alternative to Arm, whether or not Arm wins.

As we pointed out before, however, Qualcomm's interest in RISC-V isn't new. The company started using RISC-V for microcontrollers in products in 2019, and [17]it invested in SiFive , a RISC-V chip designer that competes with Arm, the same year.

Varma says RISC-V needs improvements

While Varma spent a good deal of her keynote espousing the benefits of RISC-V, she issued a call-to-arms for the development community around the ISA, saying contributors need to work on standardizing features and reducing fragmentation, a known issue among developers.

Arm shells Qualcomm's Snapdragon launch party with latest salvo in license war [18]READ MORE

"As a silicon vendor, we see a need for standardized, RISC-V-compatible system IPs. We see a lot of innovation in the application processor space. We see a lot of diversity, competition, differentiation, and that is great. But we need to make sure that we standardize the system IPs to reduce ecosystem fragmentation," Varma said.

The Qualcomm executive said the RISC-V community needs to ensure it doesn't regress on features that are being marketed for the ISA.

Varma added other wish list items for RISC-V:

We need a competitive compiler optimizing RISC-V instruction set architecture. An open-source kernel community that is picking up the latest and greatest instruction set architecture features. Standardizing a software security stack that is compatible with RISC-V instruction set architecture is essential. We can really accelerate the adoption of RISC-V software by focusing on tools, libraries, and languages.

If the RISC-V community developed a "best-in-class architecture specification" for security, machine learning, and AI, and introduced new features at a faster clip, the ISA's ecosystem "would have a tremendous time-to-market advantage," Varma concluded. ®

Get our [19]Tech Resources



[1] https://riscvsummit2022.sched.com/

[2] https://www.theregister.com/2022/11/15/qualcomm_snapdragon_8_gen_2/

[3] https://www.theregister.com/2019/12/03/qualcomm_snapdragon_tech_summit/

[4] https://riscvsummit2022.sched.com/speaker/manju_varma.24oce5s5

[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Y5tSsQJ8vwUN6UMctGsQMAAAAIs&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y5tSsQJ8vwUN6UMctGsQMAAAAIs&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Y5tSsQJ8vwUN6UMctGsQMAAAAIs&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y5tSsQJ8vwUN6UMctGsQMAAAAIs&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[9] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Y5tSsQJ8vwUN6UMctGsQMAAAAIs&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[10] https://www.arm.com/company/news/2021/02/arm-ecosystem-ships-record-6-billion-arm-based-chips-in-a-single-quarter

[11] https://www.arm.com/blogs/blueprint/200bn-arm-chips

[12] https://www.theregister.com/2022/08/31/arm_sues_qualcomm/

[13] https://www.theregister.com/2021/03/16/qualcomm_closes_nuvia_announces_chipset/

[14] https://www.theregister.com/2022/10/14/qualcomm_arm_lawsuit_nuvia/

[15] https://www.techinsights.com/blog/editorial-arms-no-win-legal-fight?

[16] https://www.telecomtv.com/content/digital-platforms-services/as-arm-sues-qualcomm-an-open-source-rival-could-be-the-winner-45296/

[17] https://www.fierceelectronics.com/electronics/qualcomm-invests-sifive-for-risc-v-architecture

[18] https://www.theregister.com/2022/11/16/arm_qualcomm_licensing_latest/

[19] https://whitepapers.theregister.com/



word of advice for Qualcomm

EvaQ

"Do not ask what RISC-V can/should do for you, but ask what you can do for RISC-V"

Interesting quotes

heyrick

On one hand, a great benefit of RISC-V is that you can add all sorts of crap to the base instruction set...but on the other hand fragmentation is a problem (when everybody else adds crap to the instruction set and things clash).

How does Qualcomm expect to rationalise these issues without basically making themselves a gatekeeper and, in the process, turning into a mini-ARM, only without the core IP to back them up?

Re: Interesting quotes

tangentialPenguin

I've always thought that standards bodies need to come up with a kind of guide book. When it's just treated like git commits the big players always end up bending the common practices to fit their mode.

Re: Interesting quotes

StrangerHereMyself

They won't. They'll simply keep adding extensions that suit their needs and expect the entire value-chain (compilers, software libraries) to support them.

These big companies only pay lip service to standardization. The article mentions that Qualcomm needed specific extensions for its own needs, but they fail to elaborate what they were and why they needed them or why they couldn't have used other extensions. My guess is it just happened to work out for them financially or lead-time wise and they went with it, not caring whether anyone else had dome something similar.

Wow! Talk about sour grapes

Will Godfrey

This sounds more like a 10yo being caught nicking an extra scoop of ice cream!

Wording

StrangerHereMyself

I love how Qualcomm references ARM as the "legacy architecture" as if they already have one foot in the grave.

OTOH I do believe RISC-V will quickly (within 5 years) surpass ARM in the microcontroller business and the high performance SoC market shortly afterward.

The misnaming of fields of study is so common as to lead to what might be
general systems laws. For example, Frank Harary once suggested the law that
any field that had the word "science" in its name was guaranteed thereby
not to be a science. He would cite as examples Military Science, Library
Science, Political Science, Homemaking Science, Social Science, and Computer
Science. Discuss the generality of this law, and possible reasons for its
predictive power.
-- Gerald Weinberg, "An Introduction to General Systems
Thinking"