Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ARM’s (and Acorn’s) founders cried wolf when ARM was bought out by Softbank. Everything that has transpired since suggests they might have been onto something. I don’t see Nvidia being a good custodian of ARM’s IP.

Any tech company would be a better IP custodian than Softbank. I'm sure Masa was at least aware that the Allen Wu / Arm China debacle was possible but was willfully blind. As an institutional investor he sees ROI a bit different than a company that has to extract it from products and services over a longer time horizon. There's a reason why most US tech companies will not agree to China's JV requirement for access to their market.
 
Well, this was inevitable considering how much Apple hates Nvidia (who bought out ARM) over its graphics chip debacle while back. They do not want to pay Nvidia a single cent going forward and going with open source method will likely bring more chip innovation for the company.

The UK Competition and Markets Authority has ordered a further investigation after flagging concerns with the takeover, coupled with the European Commission also checking into it... it still may not happen.
 
I applied to the requisition.

Albeit, this is not the first time I have applied to a job at Apple. Even with an internal referral, the best I have ever managed over the years was a phone screening, once. Precisely 0 offers. I will cross my fingers that this application goes a bit better I suppose!

Having written as much, I already have a stack of RISC-V books from over the years and hardware too. I was stoked that OpenBSD is running on the HiFive Unmatched and that the riscv64 port is now considered officially supported, and I am hoping that the BeagleBoard Foundation will email me about production BeagleV units (preliminary samples have already been sent out to developers, but alas, I am not one of the "cool kids" I guess) given that I had much better experiences in years past with a BeagleBoard Black than I ever had with any Raspberry Pi variant. I am also looking forward to the betrusted.io Precursors I preordered last year to begin shipping next year, which though FPGA based, use a RISC-V softCPU.

Linux, GCC, golang, (O)KL4, FreeBSD, LLVM also support RISC-V targets too. In other words: a lot of the underlying tooling used by many other higher level languages and frameworks, is already primed and ready to go! I think that is exciting stuff!

As others pointed out, Micro Magic already demonstrated a 5GHz RISC-V CPU consuming merely 1W of power in 2020. I did not see mention of the 24MHz RISC-V implementation announced by ONiO.zero some time before which uses energy harvesting techniques so that it has no active power draw. Yes, there are some ARM designs which purport to use similar techniques too, but those are IP encumbered, and given how the ARM acquisition from SoftBank to NVidia seems to have stalled out, I do not take that as a good indicator for ARM based designs for others moving forward. I also did not see mention that Intel has made headlines this year for purporting to be working on their on RISC-V design, purportedly Intel even made an offer to buy SiFive, though Intel was already an early investor and I haven't seen any other headlines about such things in months.

Not to mention that while the SiFive HiFive Unmatched began shipping earlier this year which uses the "world's fastest" (I guess ignoring the Micro Magic RISC-V CPU demonstrated at 5GHz after the HiFive Unmatched and Freedom u740 cores were announced) Freedom U740 RISC-V CPUs, that is using a 28nm process from TSMC. Earlier this year, SiFive and TSMC announced that they had already created a proof of concept RISC-V iteration using TSMC's 5nm process, and I know I read elsewhere that TSMC was beginning to spin up their 2nm fabrication plant, so presumably by the time that is ready, there will be even more advanced RISC-V designs ready to take advantage of it!

Some of the other things I read in this thread kind of made me scratch my head, but maybe I have been reading too much RISC-V stuff over the years?

From my vantage: ARM is a 1980s vintage CPU ISA, and proprietary. Sure, that is great for Apple that they apparently have a perpetual license, but that is not so great for educators and implementors. Yes, M1 Apple Silicon is 64bit these days, but that is sort of like AMD64 grafts to Intel's 1970s x86 CPU ISA, rather than part of the core design. Similarly, RISC-V is from 2010 and jettisons lots of old cruft that simply is not really meaningful or applicable (I think of it as the CPU ISA equivalent to ditching floppy disk drives because now we have USB flash sticks). Vector instructions, are a default in RISC-V for example, rather than an afterthought (e.g. MMX, Neon, SSE in other CPU ISAs). Though most RISC-V cores today are 64bit, there are provisions for 128bit addressing already, and 32bit iterations, and even *gasp* some unusual 8bit and 16bit RISC-V inspired iterations (officially, RISC-V is just 32bit, 64bit and 128bit) are out there too! The fact that the entire CPU ISA is libre/free and open source, means that inventive and imaginative minds have been doing all sorts of neat things with it, I think that is a killer feature personally.

From an educators' vantage, it is wonderful that the whole thing is libre/free and open, with a lot of course work available freely online in addition to quality textbooks. Moreover there are other unusual things such as MIT's Xv6 (https://pdos.csail.mit.edu/6.828/2019/xv6.html) is sort of a re-imagining of "Lion's Commentary on Unix" only using RISC-V instead of the IP encumbered AT&T Unix code that Lion's book was based upon. That is just one of many examples of how academia has been shifting attention to RISC-V. Not to mention, RISC-V's provenance from UC Berkeley, which begat BSD. It would be an understatement to state that macOS and iOS are BSD derived. When I consulted for iXSystems circa 2013, jkh (Jordan Hubbard) had just come on board as their CTO. Prior to that, Jordan was Director of Unix Technologies at Apple for a dozen years (apparently he felt that the culture had shifted a lot after Steve Jobs passed away, and for those who don't know: iXSystems was previously known as: BSDi, and for those who also don't know: jkh founded the FreeBSD project. He also wrote macports [previously known as darwinports]) FreeBSD and macOS are pretty heavily interconnected historically speaking. In more recent years, I think that OpenBSD code branches have been borrowed from heavily as well in macOS (e.g. LibreSSL, OpenSSH).

Western Digital has their SweRV implementation, Seagate supposedly has also begun to use RISC-V, there are also interesting RISC-V iterations such as Olof Kindgren's SERV, which he's been able to utilize to cram as many as > 5000 RISC-V cores into some higher end FPGA development boards. Even Microsoft has some RISC-V related publications going back to at least 2018.

I think if you have paid attention to NVidia's public research, even before the ARM acquisition attempt, they were already publicly discussing using 64bit RISC-V cores to replace the 32bit ARM cores in their GPUs (e.g. Joe Xie's presentation from 2016:
) and they have continued to present more recent RISC-V related research, some of it looks as if it is related to NVidia's aquisiton of Mellanox (now known as NVidia Networking, which is shipping 400Gbps Infiniband implementations for example, and Mellanox was already shipping 200Gbps ethernet implementations before they were acquired). My main guess with NVidia's attempted acquisition of ARM is that it might have reduced their operating expenses as far as licensing, but who knows. I don't work for NVidia though I did interview with them in 2016 so at least they invited me to their campus, which is more than I can claim any Apple hiring process ever managed as far as my past attempts at applications.

This is just me rambling about some of the highlights of interesting RISC-V developments. I think the GaNext Project (which is a RISC-V implementation in Gallium Nitride instead of silicon, or 1980s Cray vintage high end Gallium Arsenide circuit designs) is also fascinating, and I am sure there is a lot of other stuff I am not mentioning or may not even be hip enough to know about yet.

Regardless, Linux, FreeBSD, OpenBSD, (O)KL4, GCC, LLVM tend to have code which touches a LOT of things downstream whether or not consumers realize it, if those projects already think that RISC-V is worth having their constrained libre/free open source software developer time looking at it, then it stands to reason that it is probably worth some larger commercial entities also investing some resources. Apple isn't exactly first to the party, but they don't seem to be last either, which candidly, seems kind of par for the course for them, they've always been more of a "slow and steady" consumer friendly company rather than a bleeding edge research entity. At least from my vantage, while I did buy an M1 Apple Silicon Mac last year, and it has been nice enough and all, I was really hoping that a company with as much cash as Apple could have leapfrogged everyone and gotten a RISC-V based laptop to market first. As it stands, I am guessing the team behind the PineBook Pro might beat everyone else, since they already have the chops for a $200 64bit ARM laptop, and they have another RISC-V based product at the moment (albeit, it is a soldering iron. o_O) and it seems as if there is sufficient demand for their PineBooks that they are backordered! Imagine that!
 
Last edited:
  • Like
Reactions: amartinez1660
Product ASP's are set by market supply and demand not BOM costs. Management has a fiduciary duty to maximize NPV for their shareholders.
 
I doubt Apple would go for anything open source. Most probably, it will fork out some promising open source effort and make it proprietary and keep developing it. More lock in.
 
  • Disagree
Reactions: jdb8167
RISC-V is nowhere close to being used in high-performance personal devices. It’s a great teaching platform and it has uses in low-end devices, but that’s about it. There is a lot of work to be done before it becomes a viable alternative to ARM64. But I could see Apple using it for secure coprocessors or something like that.

Also, Apple has an unlimited architecture license agreement with ARM. If they switch architectures again, it won’t be before 10-15 years.



The deal is not through yet though.
People said the same thing about Arm 3yrs ago. With enough money and engineers apple could yet again set the standard.
 
  • Like
Reactions: amartinez1660
Sources from the industry claim that Apple had its hand in developing ARM64. They were the first to ship a working ARM64 CPU and a lot of changes in ARM64 are clearly targeted at making deep out of order CPU designs (just like what Apple had in mind) fly. There is no doubt that Apple started developing its CPU IP before ARM64 was officially announced and its very likely that there was significant information flow between Apple’s CPU architects and ARM’s ISA designers. Don’t forget that Apple was one of the original founders of ARM. Connection between these companies go way back.
Yes, but Apple is famous for abandoning the technologies they helped to develop. Guess who was one for the strongest supporters of OpenCL? (and I mean OpenCL, not OpenGL). Guess who developed ObjectiveC? Guess who developed the MagSafe?... I could continue for hours...
 
  • Like
Reactions: Macbookprodude
If ARM is so great, why have they been keeping us on x86 all this time since like the 80's.

Pretty much everyone has joined the RISC-V camp after the tech/trade war began, especially China. Western companies don't want to be limited in their potential customers. Alibaba, Loongson, Red Hat, even MIPS has gone RISC-V.

Even though the NVIDIA acquisition of ARM will probably never complete, Apple is smart to begin investing in RISC-V.

My understanding is the RISC-V thing is not meant to be a computer main CPU its used more for smaller stuff like an encryption chip or the monitor maybe, stuff like that but not an M1 or Ryzen replacement. Its more or less really a an educational CPU to test and try things on. Nvidia I believe already has it somewhere in their GPU cards...I think...

If you or someone else know more than I do I will gladly listen...
 
People said the same thing about Arm 3yrs ago. With enough money and engineers apple could yet again set the standard.

People were commenting on lack of high-performance ARM chips and voicing their skepticism regarding Apple‘s ability to design such a chip. Here I am talking about something different. RISC-V still does not have core components like SIMD or virtualization approved. And of course Apple could develop their own proprietary extensions to fill in the gaps, but what would be the point? It won’t be RISC-V, but rather Apple‘s ISA based on the reduced RISC-V core, requiring them to develop tools and frameworks to work with it. Sounds like a lot of hassle to replicate what they already have with ARM.

Yes, but Apple is famous for abandoning the technologies they helped to develop. Guess who was one for the strongest supporters of OpenCL? (and I mean OpenCL, not OpenGL). Guess who developed ObjectiveC? Guess who developed the MagSafe?... I could continue for hours...

Technologies are transient, and I think it makes perfect sence to move on when’s technology has outlived its usefulness. Yes, Apple was the original author of OpenCL, but that initiative has successfully dies thanks to Nvidia and flawed committee politics. ObjectiveC is almost 40 years old now and is in a dire need for a revamp. It was not fulfilling the needs of the ecosystem anymore

ARM64 instead is a new technology that is probably as close to an optimal ISA a register-based architecture can get. It’s only now reaching its critical momentum, and Apple is spearheading the efforts. Why would they abandon it now? There always has to be a reason, some benefit from choosing a different route. I don’t see any in this case.
 
  • Like
Reactions: amartinez1660
If ARM is so great, why have they been keeping us on x86 all this time since like the 80's.

Because the high-performance personal and server computer market was monopolized by x86. Software compatibility and cost are huge factors. Transition to ARM only basement feasible today because:

- Intel has completely missed the mobile market, while ARM ltd. managed to take the opportunity and proceeded to dominate it. This meant large amounts of cash for developing ARM-based CPUs and software compatibility (via improved tools, coding practices and patches)

- ARM64 (Aarch64) was developed basically from scratch to account for best CPU design practices learned in the last 30 years. It’s a great architecture for designing fast, energy efficient CPUs. In contrast, x86-64 is mostly a set of extensions that builds on the venerable platform and the legacy cruft is probably one of the reasons why x86 has difficulty scaling

- Apple had massive amounts of cash to poach talent and build up an impressive CPU design team that were able to come up with advanced technology that nobody thought would be possible. This in turn has shown what is possible and has inspired other designers and startups.
 
Not really surprising. I think Apple is always investigating options. How long had Mac OS been running on Intel x86 before the switch. I’m guessing they’ve had R&D efforts on about all conceivable options.

It’s interesting that Apple called their new CPU’s “Apple Silicon” and “M1 / Mx” instead of “Apple ARM”. I know “Apple ARM” or similar would fit their marketing style but more than that I think they are trying to abstract themselves from labeling an architecture to the consumers. Consumers really shouldn’t know or care much if it is ARM or RISC-V, etc under the hood.
 
Once again, these savings won’t get passed on to us customers, prices will continue to go up. 🤦‍♂️
But the product WILL get better. Apple is a premium product (not a luxury product), but is still a tremendous value to those who budget appropriately for their products.
 
  • Like
Reactions: I7guy
I think Steve Jobs was bigger on retribution. Apple of today just decides they will no longer rely on or partner with the company.

Apple is still rely on retribution as part of preserving Steve Job’s legacy. The members of the group that’s part of App Fairness cartel? Apple already subpoenaed them as part of the lawsuit against Epic. And Apple may kick them off the developer program and ban them from owning any Apple products for life after the lawsuit is over as payback.
 
typical with big tech like Apple.. One time not so far in past, Apple was called Apple Computer where it cared. Now they don't care for no one.
 
Apple is still rely on retribution as part of preserving Steve Job’s legacy. The members of the group that’s part of App Fairness cartel? Apple already subpoenaed them as part of the lawsuit against Epic. And Apple may kick them off the developer program and ban them from owning any Apple products for life after the lawsuit is over as payback.

Apple didn’t subpoena them - their lawyers did. And if they didn’t, they‘d be committing malpractice. If there is a scheme to team up against apple, and Epic’s lawsuit is part of it, that’s obviously highly relevant.

Your last sentence is meaningless - “apple may.” They haven’t. If they had, that would be retribution. (And they can’t ban anyone from owning apple products for life). You can’t say “here’s proof of Apple‘s retribution. They could theoretically do something!”
 
Yes Apple is more bureaucratic
 
Once again, these savings won’t get passed on to us customers, prices will continue to go up. 🤦‍♂️

Current Canadian prices for Mac Minis
  • Mac Mini Apple M1 Chip with 8-Core CPU and 8-Core GPU 512 GB Storage, 8GB of RAM: $1149
  • Mac Mini 3.0GHz Intel Core i5 6-Core Processor with Intel UHD Graphics 630 512 GB Storage, 8GB of RAM: $1399
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.