This just sounds all kinds of wrong. NVidia I could see, but AMD?
1 Like
Okay, a second source stating that Nvidia is definitely going this way and suggesting “oh, yeah, by the way AMD is also going to make Arm CPUs according to two people familiar with the matter”
Nvidia has quietly begun designing central processing units (CPUs) that would run Microsoft’s (MSFT.O) Windows operating system and use technology from Arm Holdings(O9Ty.F), , two people familiar with the matter told Reuters.
The AI chip giant’s new pursuit is part of Microsoft’s effort to help chip companies build Arm-based processors for Windows PCs. Microsoft’s plans take aim at Apple, which has nearly doubled its market share in the three years since releasing its own Arm-based chips in-house for its Mac computers, according to preliminary third-quarter data from research firm IDC.
Advertisement · Scroll to continue
Report this ad
Advanced Micro Devices (AMD.O) also plans to make chips for PCs with Arm technology, according to two people familiar with the matter.
Nvidia and AMD could sell PC chips as soon as 2025, one of the people familiar with the matter said. Nvidia and AMD would join Qualcomm (QCOM.O), which has been making Arm-based chips for laptops since 2016. At an event on Tuesday that will be attended by Microsoft executives, including vice president of Windows and Devices Pavan Davuluri, Qualcomm plans to reveal more [details about a flagship chip](https://www.reuters.com/technology/nvidia-make-arm-based-pc-chips-major-new-challenge-intel-2
023-10-23/according%20to%20a%20person%20familiar%20with%20the%20matter) that a team of ex-Apple engineers designed, according to a person familiar with the matter.
Nvidia shares closed up 3.84%, and Intel shares ended down 3.06% after the Reuters report on Nvidia’s plans. Arm’s shares were up 4.89% at close.
Nvidia spokesperson Ken Brown, AMD spokesperson Brandi Marina, Arm spokesperson Kristen Ray and Microsoft spokesperson Pete Wootton all declined to comment.
Nvidia, AMD and Qualcomm’s efforts could shake up a PC industry that Intel long dominated but which is under increasing pressure from Apple (AAPL.O). Apple’s custom chips have given Mac computers better battery life and speedy performance that rivals chips that use more energy. Executives at Microsoft have observed how efficient Apple’s Arm-based chips are, including with AI processing, and desire to attain similar performance, one of the sources said.
The Reuters piece made reference to SemiAccurate having reported this earlier but I can’t find it. Part of their subscriber only content, I guess?
Did they know/hint two months ago that this was coming? Anyone out there have semiaccurate paid access?
1 Like
First, on the difficulty, or lack of it, for AMD to start making ARM chips again. (They did it about a decade ago, and still have ARM code as part of their hardware protection logic for EPYC.) So, the licensing part should be easy to complete. As for design, take current chip designs and replace the instruction decode logic. Since this is a deterministic finite-state machine, there are tools to convert those to logic gates. I could write that part as a context-free grammar that outputs x86 instructions or AMD’s RISC instructions. There might also be some changes to the termination logic for instructions. x86 requires in-order completion of instructions. Not having that requirement would result in additional dependencies between instructions, but those would be handled explicitly in the DFSM. The tricky part will be the interrupt logic. The x86 rule is nice in that the interrupt state only has to remember the last instruction completed. Yes, there are many interrupts where there is no need to save anything. For example, a clock interrupt where there are no threads waiting on that tick.
Most of you probably found that interesting but had no idea what half of it meant. So, simplified translation: AMD could reuse all the (incredibly) hard parts of x64 chip design and just change some of the easy parts to create a fast ARM chip. In fact, AMD could easily add the circuitry to current chips, and mix ARM and AMD64 code.
Now, about the rumor. I put no faith in it at all. This is the sort of “rumor” that some floor trader or specialist starts when they have stock to sell. It might create just a short bump in the price of the stock, but that is all that is needed.
I keep thinking that the SEC should put a stop to this. If it requires putting enough microphones on the trading floors so that they can sort out what every individual is saying, fine. Nothing can be said there with an expectation of privacy. Maybe add a few cameras to make tying voices to people easier. But that is not what is needed. The SEC has access to both sides of every transaction. If a rumor starts and someone who may have started it benefits, that’s all you need. The SEC should probably start out with a hand slap, then get very serious if the floor trader or brokerage firm gets caught red-handed again. Then, it is time to start with suspensions or worse.
Now, about the rumor. I put no faith in it at all. This is the sort of “rumor” that some floor trader or specialist starts when they have stock to sell. It might create just a short bump in the price of the stock, but that is all that is needed.
Yeah, to your technical point I’m sure they’re capable of doing an Arm chip, I’m just hazy on why they would do it. My only thought was if you want a literal all-day 12-16 hour laptop life in a super light weight MS Surface Tablet form factor, maybe it’s impractical to get there even with e.g. Zen 5c cores, so to stay in the game you start to offer an Arm-based very low power alternative. (No idea whether that’s where NVidia is trying to get to with any hardware they’re making.) Maybe in so doing you can sidestep some compromises you’d otherwise have to consider in Zen 6 to make it suitable for the full range of laptops to high end servers. Zen 4c and 5c are a clever, economical use of available techniques to avoid designing a separate E-Core and are a good way to help servers draw less power but maybe that’s not enough for the mobile market.
Or maybe there’s some major advantage to going away from x86 to help your marketing partner Microsoft, or simply some way in which non-x86 lets you elude some licensing fee to Intel by letting you grow sales outside the cross-licensing agreements etc.
We wait. Earnings is in a week, I think. If there’s anything to say one way or the other, it’ll be said then.
Adapting GPU chips to run Windows probably means much faster processing. Traditionally that allows more sophisticated software.
Plus Intel is the leading supplier. You take a shot at grabbing market share.
As AMD makes chips that run Windows already, its a potential conflict for them. Presumably these will be premium performance chips at premium prices. Its a no brainer for Nvidia. Whenever they have the resources. Nothing to lose and lots to gain.