Slash Boxes

SoylentNews is people

posted by martyb on Tuesday May 04, @05:47AM   Printer-friendly [Skip to comment(s)]
from the Arm-ageddon dept.

Arm Pioneer: Nvidia's Grace CPU Is Proof That It Will 'Compete Unfairly'

Arm pioneer Hermann Hauser has once again criticized Nvidia's plan to acquire the semiconductor design company, with The Telegraph reporting Sunday that he believes Nvidia is "clearly showing it will compete unfairly" if the deal is approved.

Hauser's concerns reportedly centered on the Grace processor Nvidia announced at GTC 2021. The company's first Arm-based CPU will connect to high-end GPUs via NVLink, which purportedly offers data transfer speeds up to 900 GBps. That's significantly faster than other technologies—it's also exclusively available to Nvidia.

This is why Hauser told The Telegraph that he believes using a proprietary interface like NVLink could end up "locking customers into [Nvidia] products," which "clearly shows that they will compete unfairly with other Arm-based server companies such as Amazon and Fujitsu," rather than retaining Arm's neutrality.

Nvidia-Branded ARM CPUs; UK Trade Union Speaks Out Against Deal
Nvidia's $40 Billion ARM Acquisition: "All but Dead"?
Nvidia Enters the Arms Race with Homegrown "Grace" CPUs
UK Will Investigate Nvidia's $40 Billion Arm Acquisition Attempt

Original Submission

Related Stories

Nvidia-Branded ARM CPUs; UK Trade Union Speaks Out Against Deal 9 comments

Jensen Huang Says Nvidia-Branded ARM CPUs Are a Possibility

According to comments from Nvidia CEO Jensen Huang during a conference call yesterday, we could see Nvidia-branded CPUs in the future, setting the stage for a new level of competition with Intel and AMD.

[...] However, during yesterday's briefing, Timothy Prickett Morgan from TheNextPlatform asked Jensen Huang, "Will you actually take an implementation of something like Neoverse first and make an Nvidia-branded CPU to drive it into the data center? Will you actually make the reference chip for those who just want it and actually help them run it?"

"Well, the first of all you've made an amazing observation, which is all three options are possible," Huang responded, "[...] So now with our backing and Arm's serious backing, the world can stand on that foundation and realize that they can build server CPUs. Now, some people would like to license the cores and build a CPU themselves. Some people may decide to license the cores and ask us to build those CPUs or modify ours."

"It is not possible for one company to build every single version of them," Huang continued, "but we will have the entire network of partners around Arm that can take the architectures we come up with and depending on what's best for them, whether licensing the core, having a semi-custom chip made, or having a chip that we made, any of those any of those options are available. Any of those options are available, we're open for business and we would like the ecosystem to be as rich as possible, with as many options as possible."

Nvidia's $40 Billion ARM Acquisition: "All but Dead"? 18 comments

Nvidia's US$40 billion deal to buy Arm is all but dead – it's a classic example of geopolitics killing innovation

Under normal circumstances, US tech giant Nvidia's takeover of British chip designer Arm for US$40 billion (£29 billion) would have sailed through without registering beyond the computing industry. Instead, it has made international headlines, with UK and EU monopolies regulators launching an in-depth investigation after outcry from competitors.

In effect, the deal is pretty much dead before it starts. At the heart of this lies a row about technological sovereignty. So what is going on?

[...] The biggest pushback, behind the scenes, actually appears to be from China. Ever since the US blacklisted Huawei and other semiconductor manufacturers in China, Beijing has been obsessed with becoming technically "self-sufficient".

While it works towards this goal, Arm has continued to license its chip architectures to Huawei. Arm claims that its chip technology is of British origin and therefore does not breach the US restrictions on exporting tech to a group of blacklisted Chinese companies. Thanks to this ongoing arrangement, Arm is one of the remaining enablers for China's semiconductor sector to keep pace with the outside world.

See also: ANALYSIS-Nvidia acquisition of Arm throws company into tech spat between U.S. and China

Previously: Nvidia Announces $40 Billion Acquisition of Arm Holdings
Nvidia-Branded ARM CPUs; UK Trade Union Speaks Out Against Deal

Original Submission

Nvidia Enters the Arms Race with Homegrown "Grace" CPUs 13 comments

Nvidia Enters The Arms Race With Homegrown "Grace" CPUs:

Everything that Nvidia and the market has learned about CPUs in the past decade can be brought to bear in Grace CPUs, and now Nvidia itself is big enough to be the primary contractor for exascale-class machines or at the very least supply the CPU and GPU compute as well as the networking and DPU offload engines to a supplier such as Hewlett Packard Enterprise, which really doesn't mind making custom variants of its "Shasta" EX supercomputers. And that is precisely what will be happening with the future "Alps" supercomputer going into the Swiss National Supercomputing Center [(CSCS)] in early 2023, featuring a hybrid compute module that has the homegrown Grace Arm server CPU very tightly couple with a future Nvidia GPU accelerator.

[...] CSCS is not alone in getting a Grace-Hopper supercomputer. Los Alamos National Laboratory, one of the Department of Energy national labs in the United States, will also be getting a Grace-Hopper system supercomputer system HPE in early 2023.

[...] Nvidia estimates that a Grace-Hopper system will offer 10X the performance on training natural language models, reducing it from one month to three days, and will allow for real-time inference on a 500 billion parameter model on a single node. And on the HPC front, computational fluid dynamics and weather applications are expected to see performance on the Grace-Hopper systems as well. Other HPC applications will no doubt follow.

See also: For Arm-Driven Supercomputing, Nvidia is Right on Time
Intel CEO: 'We Are Playing Offense, Not Defense' In Response To NVIDIA Grace ARM CPU Announcement For Servers

Original Submission

UK Will Investigate Nvidia's $40 Billion Arm Acquisition Attempt 22 comments

Nvidia's Share Price Dips as UK Goverment Questions $40 Billion Arm Acquisition

The UK Secretary of State for Digital, Culture, Media and Sport today announced that he is asking the Competition and Markets Authority (CMA) to investigate the Nvidia-Arm merger because he believes the deal could have national security implications.

CMA has until July 31 to "prepare a report on the competition and national security aspects of the proposed transaction" for the UK Secretary of State. Then it's up to the Secretary to decide if Nvidia's acquisition of Arm will be allowed to proceed.

Also at CNBC and Bloomberg.

Previously: Nvidia Considering Acquisition of ARM for Over $32 Billion
Nvidia Announces $40 Billion Acquisition of Arm Holdings
Nvidia-Branded ARM CPUs; UK Trade Union Speaks Out Against Deal
Nvidia's $40 Billion ARM Acquisition: "All but Dead"?

Original Submission

Display Options Threshold/Breakthrough Reply to Article Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Insightful) by Anonymous Coward on Tuesday May 04, @06:07AM (6 children)

    by Anonymous Coward on Tuesday May 04, @06:07AM (#1146010)

    how is the use of nvlink unfair?
    are they forbidding others from making their own am-based derivatives, that use different connections?

    if you don't like the fact that nvlink is proprietary, then ask for more sensible legal handling of the ridiculous "intelectual property" concept.
    but, unless arm-derivatives are subject to a gpl-like opennness requirement, nvidia is perfectly right to make use of its proprietary tools.
    same as apple making use of bsd.

    • (Score: 4, Interesting) by RamiK on Tuesday May 04, @10:48AM (2 children)

      by RamiK (1813) on Tuesday May 04, @10:48AM (#1146073)

      Interlink isn't just an added logic block you can simply go to a third party and license if you don't like Nvidia's offer. It's part of the whole design (lane number, speed and protocols) and can be used by nVidia to segment the market and prevent competition at the same time by simply telling ARM they're not to offer a solution from their own labs and force them to design per NVLINK's needs.

      And it's not unprecedented within ARM: The reason Qualcomm and Apple are designing their own ARM-based cores is precisely because ARM is bundling all sorts of under-performing stuff (GPU, power governing, wifi, baseband...) with their base licenses in hopes you'll license their high-end stuff instead even when there's no technical reason not to use the better designs (power/performance/BoM on all use cases are superior...).

      • (Score: 0) by Anonymous Coward on Wednesday May 05, @02:41PM (1 child)

        by Anonymous Coward on Wednesday May 05, @02:41PM (#1146466)

        who cares! ARM is not in the standard chip-making business! every ARM licensed to samsung, apple etc is it's own little stamp with its own design.
        it's a freaking ABOMINATION that a vanilla standard ARM-! linux kernel cannot boot on all and each of them 3 year throw-away devices!
        go cry me tears! i hope ARM rots away on the seafloor in some "oops fell overboard mobile selfy phone"!

        • (Score: 0) by Anonymous Coward on Wednesday May 05, @03:25PM

          by Anonymous Coward on Wednesday May 05, @03:25PM (#1146476)

          "yes sir!"
          "you are going to ARM today to get new designs for the new phone chip!"
          nervous:"Do i have to, boss?!"
          "yes! yes! be sure to also get a appointment with their 'special division'. but don't mention them by name!"
          "okay ..."
          later that day at ARM headquarter
          "Hello Mr. Park! long time no see! how are things?"
          "good good ..."
          "follow me to conference room and we can get down to business..."
          "so you would like the new chip designs? they are very good indeed!"
          "yes, this is why am here"
          "we have new options for .." and here a long list of tech specs rambles along
          "yes, it looks very good. we will take it, as always but we would like some ..uhm ..errr ..modifications?"
          "of course, of course. we can help with that" big smile
          "oh i am relieved ..."
          "what "modifications" *wink* "did you have in mind?"
          "well, we only want the best for our customers and some of them feeel very attached to their devices after some time and we ..."
          "say no more, sir" writes down some number and address "here is the phone number of our special division; we call it the '1095 division'. the number is for mr quirk, the head engineer!"
          "yes, he can help you with, the .. uhm ..errr ... 'attachment problem'. he will help you find appropriate programmers well versed with the linux kernel. the modifications will guarantee no extra benefit and no extra cost to your ARM licensed chip design but will ..uhm ...errr ... help with the 'attachment problem'!"
          "this is wonderful! my boss will be very happy to hear this and the designed-in obsoletness will surely make us come back to you again!"
          "now now mr.park, we don't call it that!"
          "of course ..."
          "so it's settled then? anything else? maybe some extra battery life"
          "no, no, that's all" wipes sweet from forehead.
          "thank you very much mr. ARM, see you in 3 years!"
          "see you in 3 years mr. park!"

    • (Score: 5, Interesting) by theluggage on Tuesday May 04, @10:56AM (2 children)

      by theluggage (1797) on Tuesday May 04, @10:56AM (#1146076)

      how is the use of nvlink unfair?
      are they forbidding others from making their own am-based derivatives, that use different connections?

      I don’t think TFS made the problem very clear. You’re quite right that others are free to add proprietary tech to their own ARM chips (Apple Silicon is made of that) but they don’t own ARM.

      ARM don’t make and sell chips: they license designs. If they design a new CPU core, GPU, interface, whatever, their only business plan is to license it to others so it becomes part of the “standard” ARM ecosystem. They have no incentive to design stuff and keep it to themselves. That’s not theoretical: they already have a catalogue of such designs. That model has been a key part of ARM’s success, as it lets the smaller players (who don’t have Apple, Amazon’s, etc capacity to build processors from scratch using just the instruction set) build their own pick’n’mix processors.

      If Nvidia buy ARM then they will have *every* incentive to plough any investment into proprietary extensions that make “their” ARM chips better than others, and the “standard” ARM component catalogue could be left to stagnate.

      Having ARM owned by a chipmaker who will prioritise their own products over tech licensing won’t be good for ARM. Whether that is a legal justification for blocking the takeover is one for the lawyers.

      • (Score: 2) by crafoo on Tuesday May 04, @11:11AM (1 child)

        by crafoo (6639) on Tuesday May 04, @11:11AM (#1146081)

        Insightful, also, such an acquisition of ARM may also provide incentive to spawn new RISC CPU design and licensing companies that would have a better chance of competing with a locked-down ARM catalog.

  • (Score: 5, Informative) by FatPhil on Tuesday May 04, @08:16AM

    by FatPhil (863) <{pc-soylent} {at} {}> on Tuesday May 04, @08:16AM (#1146049) Homepage
    "The company's first Arm-based CPU"

    A first look at NVIDIA’s APX 2500 - Page 1
    Published on 12th Feb 2008, written by Arun for Handheld 3D - Last updated: 12th Feb 2008

    NVIDIA is demoing a new in-house 3D GUI to showcase the APX 2500's capabilities.

    NVIDIA has just announced the 65nm APX 2500, an application processor supporting 720p H.264 video, OpenGL ES 2.0, and HDMI output. On the processing side, it sports an ARM11 core at 750MHz.

    The whole Tegra family was ARM-based, that model's just a rando example from the range.

    Apparently in IT 13 years is long enough to have forgotten everything, and thus for everything old to be new again.
    I know I'm God, because every time I pray to him, I find I'm talking to myself.
  • (Score: 1) by shrewdsheep on Tuesday May 04, @08:47AM (4 children)

    by shrewdsheep (5215) Subscriber Badge on Tuesday May 04, @08:47AM (#1146055)

    Why should Nvidia buy ARM? They can build custom chips right now. What could the "synergies" be? Probably not killing the licensing model as that is the only income of ARM. But certainly having access to designs early opening them up for licensing only later. This way they might force some proprietary bus on some phone makers eager to boast cutting edge chips.

    BTW I always wonder how ARM keeps being alive. ARM is a company truly running on human capital. How do they keep their designers happy not to leak all the business secrets? Or is the design factored enough so that individuals only oversee a tiny fraction of the overall design?

    • (Score: 2) by crafoo on Tuesday May 04, @11:14AM (3 children)

      by crafoo (6639) on Tuesday May 04, @11:14AM (#1146083)

      It could be simply that they offer a better employment deal than any other company + whatever another company would be willing to pay for the experience and knowledge. This happens in other engineering fields, where experienced engineers make $$$bank (eventually). Usually MBAs come in, get cold sweats and pale faced when they see the payroll #s, then start force-retiring the older guys. These guys go on to start a new company that eats their old employer's lunch. It's the cycle of life in engineering.

      • (Score: 5, Insightful) by PiMuNu on Tuesday May 04, @11:35AM (1 child)

        by PiMuNu (3823) Subscriber Badge on Tuesday May 04, @11:35AM (#1146085)

        Or just don't irritate the engineers. It turns out that many (most?) people don't want to take on risk and hassle of moving jobs (houses, schools, etc etc). Pay them enough, don't irritate them. Let them or even encourage them to have fun. That's all it takes.

        Maybe the chase of $$$ is an American thing?

        • (Score: 1) by shrewdsheep on Tuesday May 04, @12:29PM

          by shrewdsheep (5215) Subscriber Badge on Tuesday May 04, @12:29PM (#1146103)

          Probably that is important. One can easily target key employees and outbid the previous employer by whatever it takes. Everyone can be bribed, it is just a matter of amount. OTOH Lisa Su could not be lured away from AMD.

      • (Score: 0) by Anonymous Coward on Saturday May 08, @10:22PM

        by Anonymous Coward on Saturday May 08, @10:22PM (#1147623)

        It used to work that way back when the engineers owned their own patents. Even if the engineers left they would still pay the patent holder a licensing fee to use the patent. Now companies own the patents (and the patents now a days are stupid anyways).

        Perhaps a solution is to make it illegal for a company to require patents to be transferred to them as an employment term. The individuals that came up with the invention own the patent and they can't be fired for refusing to transfer over the patent and the company can not include any terms in an employment agreement that requires patents to be transferred over to them. It can be a type of labor law, companies can't require inventors to hand the fruits of their labor over to the company. If the company wants perhaps it can buy or license the patent from an individual if they are willing to offer enough money to persuade them but that's it. Any indication that the company otherwise discriminates against those that refuse to transfer or sell over their patents in any way shape or form should be breaking the law.

        If the employee leaves their patent can go with them if they never sold it. Or the employer can offer to buy it from them. Or the employee can license it to multiple companies (they used to back in the days) or decide to sell it to a competitor even.

        Perhaps the company should also be allowed to require that the individual patent holder must indefinitely license them any patent that they acquire while on company time for some reasonable licensing fee (the law can determine an upper and lower bound and the terms of employment can specify the exact amount). Even if the patent is later sold to another company that required licensing must also get transferred and disclosed before any patent sales are made. That way a company isn't stuck with their entire business being in the hands of a few patent holders either.

  • (Score: 0) by Anonymous Coward on Tuesday May 04, @07:59PM

    by Anonymous Coward on Tuesday May 04, @07:59PM (#1146278)

    Let's level the playing field, so there is no unfair advantage for anybody. That's a sure way to have famine, war and ignorance. But it will be fair. Life is not about having an ace up your sleeve, but the ability to play the game better than the other guy.

  • (Score: 1, Interesting) by Anonymous Coward on Tuesday May 04, @11:03PM (2 children)

    by Anonymous Coward on Tuesday May 04, @11:03PM (#1146319)

    Just how involved in ARM Ltd. is Hauser? He seems to have strong opinions on who should own it, but it also sounds like his big contribution to ARM was selling it off way back in the day.

    He was sad about the sale to Softbank too: []

    However ARM co-founder, Hermann Hauser who is now a partner at Amadeus Capital, said in a tweet that the deal represented a "sad day" for the UK's technology sector. []

    Hermann co-founded Amadeus Capital Partners in 1997 with Anne Glover. In his long and successful history as an entrepreneur and venture capitalist, Hermann has founded or co-founded companies in a wide range of technology sectors. These include Acorn Computers (where he helped spin out ARM) ... []

    Apple and Acorn began to collaborate on developing the ARM, and it was decided that this would be best achieved by a separate company.[79] The bulk of the Advanced Research and Development section of Acorn that had developed the ARM CPU formed the basis of ARM Ltd. when that company was spun off in November 1990. Acorn Group and Apple Computer Inc each had a 43% shareholding in ARM (in 1996),[80] while VLSI was an investor and first ARM licensee.[81] [] (Hauser is not listed as a founder or even anywhere on the ARM wiki)

    ... by February 1999, Apple's shareholding had fallen to 14.8%.

    From Hauser's wiki: []

    [he first worked with Acorn Computers, which got bought by Olivetti, where he worked on a book sized computer, which AT&T picked up, but which failed and then] Hauser was involved in spinning out Advanced RISC Machines (ARM) from Acorn.[citation needed]

    You see this "citation needed" all over the place with Hauser and his relationship to Arm. Hauser calls himself a cofounder but perhaps he was merely responsible for selling off ARM in 1990. It's unclear to me what his issue is other than that he doesn't want a non-UK company to own ARM, but no matter how he was involved and to what degree he was involved with ARM, it looks like that involvement was 30 years ago. Maybe it's time he let whatever issue is at play here, go.