Hacker News Reader

Comment by awalton | Hacker News Reader

I'm in agreement. nVidia's history of not playing nicely with others does not make this sound like a good deal, especially for entrenched ARM vendors. (Apple won't be phased; they basically design everything themselves, so they don't really even care about the direction of ARM or the architecture. They could hard fork tomorrow, or two years ago, and nobody will notice since they're entirely vertically integrated.)

But, on the plus side, it's the biggest opportunity RISC-V will ever get. Almost overnight anyone who was building ARM cores will suddenly see investments in RISC-V prudent as a hedge against nVidia's future core designs not being licenseable. SiFive will very quickly become a multi-billion dollar organization...

View on HN

Apple’s ARM chips contain an awful lot of different components, some of which are highly customised. The neural engine stuff they use for image processing for example. The CPU cores are custom tuned, but I believe they are still 90% based on ARM CPU core designs. For example the Secure Enclave is actually an ARM developed CPU extension. Apple focuses their efforts vary specifically on areas that give them features that grant a competitive advantage. They don’t reimplement stuff for the sake of it. One of their big performance wins over other ARM licensees isn’t even really in the CPU, it’s the very well designed and generous cache memory systems.

Having said that, looking to the future I think Apple is clearly willing to pour more resources into their CPU programme than any of their competitors. This means the risk to them is probably a lot less than to say Qualcomm or Samsung as they have a much better capability to go their own way. As time goes on their designs will probably diverge more and more from main line ARM anyway.

Apple‘s designs are already far better than the equivalent ARM coRe. I doubt they're using anything besides the ISA.

Apple licenses the isa only. They implement their own micro architecture.

Apple has an architecture license from ARM, which allows them to design their own cores yes, but also includes all the benefits of all the lower tiers including rights to all the ARM designs.

This would have to be the case, because Apple still uses many of their earlier ARM designs in products, which we know used ARM designed cores and would definitely require core design licenses.

All this seems fitting since they are an ARM founder.

I think nVidia could run into some serious anti-trust issues if they tried to interfere with the ARM ecosystem. There's far too much at stake here. Is there a phone on the planet that doesn't run ARM?

The Asus Zenfone 2, released 5 years ago, ran on x86 (Intel Atom) [1]. I know someone who was still using it this year Eventually switched but the phone still works. So, to the question:

> Is there a phone on the planet that doesn't run ARM?

Well, technically, yes ;-) This does not change anything to your point though.

[1] https://en.wikipedia.org/wiki/Asus_ZenFone#Second_generation...

I bought the Zenfone 2 which had an Intel processor, and had a similar experience to the sister comment. The only super annoying thing that happened was when Pokemon Go came out all my friends had it but it didn't support Intel. By the time they made it available for the intel phones (which was a few weeks) the game wasn't cool anymore. :/

I used the ZenFone 2 for about a year ~3.5 years ago; it worked surprisingly well. It was impressive that it really "just worked" like a normal ARM device.

There were some annoying software aspects (the Play Store only allowed me to download x86 apps even though it could run ARM apps), and the device was 100% a battery hog - but cool tech.

Surprised Intel gave up on that effort.

Intel gave up because Atom for smartphones were sold at a negative price after marketing incentives. They couldn’t even get volume from manufacturers when it costed less than zero; because no one wanted to invest upfront R&D costs.

Intel should have made their own phones.

There were a few MIPS Android phones too.

MIPS was pretty popular among early Shenzhen tablets.

Then they just disappeared. Either MIPS screwed up or ARM did something really smart. My money is on the former...

> MIPS was pretty popular among early Shenzhen tablets.

> Then they just disappeared. Either MIPS screwed up or ARM did something really smart. My money is on the former...

MIPS SoCs simply fell too much behind. They kept presence in the STB, and router boxes for some time, but even there they lost due to overall "deadness" of the entire ecosystem:

Want latest compiler work, and mainline GCC support? Go for ARM

Want to not to have to support your own Linux fork? Go for ARM

Want more IP coming from a single vendor that just core, interrupt, controller, and DMA? Go for ARM

Want to work with an IP vendor whose sales are not part time lawyers? Go for ARM

When MIPS understood that they are losing, they simply gave up on competing, and improving.

And in the end, the lunch went to MediaTek because it was the only SoC vendor that was ready to ship competitive SoCs with 4G baseband without rapacious licensing deals like Qualcomm do.

Actually at that time MIPS suddenly came back to life and had a big push with improved designs and a few deals resulting in pretty decent MIPS socs. Performance was okay, battery life was good, and they were first to market with the latest Android release (I think it was 4.0 or 4.4).

But then everything suddenly stopped and they left the market altogether.

Edit: Regarding this:

> Want to work with an IP vendor whose sales are not part time lawyers? Go for ARM

One of the founders of OpenCores (RIP) told me that ARM lawyers contact him _within hours_ of someone pushing a design they dont like. While MIPS has a bad rep, ARM is probably more aggressive.

> Want to work with an IP vendor whose sales are not part time lawyers? Go for ARM

Er... not convinced about this one.

ARM has been quite lawyerish sometimes.

For a while, parts of QEMU could not be published, at ARM's request, because they implemented instructions ARM did not permit a software emulator to emulate.

At least one non-commercial design on opencores.org was removed at ARM's request.

By comparison, for a while MIPS ISA was completely open for anyone to implement, software or hardware. That's been retracted now, but it was nice while it lasted. ARM has never done that as far as I know.

Even before, I used devices based around MIPS without a licence a long time ago, so long as they avoided implementing one or two "special sauce" instructions that was fine. The instructions weren't necessary, just a neat, patented optimisation, and I have a GCC patch somewhere for building the target without those instructions.

So I don't know about today, but until recently I would have consistently regarded MIPS as much less lawyerish than ARM.

> Want to work with an IP vendor whose sales are not part time lawyers? Go for ARM

Interested in this comment: were MIPS/Imagination known for being particularly litigious around licensing?

There are tons of stories about MIPS going after people for trivial designs.

http://probell.com/Lexra/index.html

> Want latest compiler work, and mainline GCC support? Go for ARM

> Want to not to have to support your own Linux fork? Go for ARM

But that's also true for MIPS.

I have not seen any activity from MIPS on GCC for years.

Making a Linux port for a given MIPS based SoC is a known pain. They should've worked towards not requiring any hardcoded configuration for each SoC at all, like Linux on ARM does these days.

Do you mean the device tree?

I thought all major kernel archs were on device trees now...

Yes now, but last time I talked with people who dealt with them, people spoke of their new cores coming with "prehistoric kernels" on usb sticks.

So every vendor had to upstream their own soc in its own

Ubiquiti will still sell you a shiny, expensive MIPS box, as will GL.inet.

Interestingly, they've gone ARM in their latest generation of hardware, e.g. the UDM-Pro uses a Quad ARM Cortex-A57 Core at 1.7 GHz: https://store.ui.com/collections/unifi-network-routing-switc...

So their MIPS stuff might be consigned to the past at this point.

>I think nVidia could run into some serious anti-trust issues if they tried to interfere with the ARM ecosystem.

I seriously doubt it, Broadcom - Qualcomm sale was blocked because there were concerns that the former would inhibit the telecom technologies development of the latter, resulting in strategic disadvantage to U.S. But, in the case of supposed Nvidia-ARM sale I see the quite opposite i.e. huge strategic advantage to U.S. and so I feel the anti-trust might not be at least State initiated.

But, would other U.S. Big Tech who have invested heavily in the ARM ecosystem oppose this is what to be seen. But even there, I wonder whether their fear could be substantiated? What if Nvidia just signs a good agreement with them to alleviate their fears? After all Apple and Qualcomm were strangling each other throats, now see where they are now.

In the end, I think the deal would just go through and it's us who are worried about the future our Raspberry Pis among others things who would feel lost(and actually loose).

Wouldn't the EU antitrust look into that too or is this going to be out of their jurisdiction after Brexit?

Even if the EU did not have jurisdiction British authorities are not very keen on large mergers either.

British resistance caused American Pfizer to walk away from a potential merger with British AstraZeneca.

I'm not sure how much sway British authorities would have, since ARM is now wholly owned by SoftBank?

In a merger of this scale all the big powers (US, EU, China, and I assume also UK now) have basically veto right.

There are countless examples where country 1 has stopped a merger between a company from country 2 and one from country 3.

> In a merger of this scale all the big powers (US, EU, China, and I assume also UK now) have basically veto right.

> There are countless examples where country 1 has stopped a merger between a company from country 2 and one from country 3.

From what I've seen, the only method of enforcement for a third-party country is to ban the sale of goods in their respective markets following the merger (like GE/Honeywell in the early 2000s). Would that even be possible in this case?

Not sure how this is enforced. Generally the merger falls apart if there are any indications of this happening.

Don't forget that these are public companies and a stopped merger can affect their stock. Besides, major investors would eat the board alive if they put the company in such situation.

Antitrust regulators power is determined by the size of the market, not the location of the company HQ.

Antitrust regulators for big markets like the US, EU can effectively block mergers from any internationally operating company. Both AMD and Nvidia need access to EU - their customers certainly do.

> Antitrust regulators power is determined by the size of the market, not the location of the company HQ.

> Antitrust regulators for big markets like the US, EU can effectively block mergers from any internationally operating company. Both AMD and Nvidia need access to EU - their customers certainly do.

They could block the merger and if the companies decided to go ahead with it anyway, the EU would ban imports. But, wouldn't that further consolidate the market?

> and if the companies decided to go ahead with it anyway

This has never happened, because the profit loss from both parties from losing access to a market with 400M+ people has always made investors blanche.

What I mean is, in what practical way could the EU ban ARM/Nvidia from their market, considering these manufacturers' chips are in many devices? Would that mean devices with those components would also be banned?

What I'm trying to get at is, with a sufficiently large merger, the emerging company may produce products so ubiquitous that it would be impractical to bar them from your economy, due to the widespread use of their products and their integration into other products.

I believe as long as ARM sells their hardware in the EU, they are able to look into antitrust issues.

Not that I disagree with you, but ARM doesn’t sell hardware—they sell licenses to designs of the aforesaid. Nonetheless, don’t forget that nVidia, insofar as it sells hardware in the EU, would definitely be within the scope of local anti-trust authorities.

I don’t see this being much liked by those authorities, either. nVidia is clearly interested in doing this deal because it will give them competitive advantages, and it’s pretty clear by their behaviour versus other firms (including Apple!) that they have so much of that already that they don’t have to “play nice”. Giving them further purview to do as they please is basically what anti-trust is designed to prevent, whether you focus on industry relations or the ascendant on consumers.

I think there’s some MIPS and SH-3/SH-4 smartphones from the early 2000s still around.

I had an SH-3 PDA in 2001, but Microsoft dropped support for SH-3 and MIPS in Windows CE by the time “Windows Smartphone 2002” came out (my 2002 HTC Canary runs ARM, IIRC) so it’d have to be a non-CE platform. What did the Palm Handspring use?

Internally we supported SH4 through CE7!

Source: I maintained the compiler test infra for it.

Not only internally. Infotainment Systems with the CE7 and SH4 combination had been built well beyond 2012. Only around that time Arm got traction.

At some point in time it got then renamed in embedded compact.

Source: worked in the automotive together with a couple of supplier which built those systems.

SH4 is out of patent now - do you know if the J-Core can run CE7?

Unlikely, since the SoC is different.

I salute you sir!

What’s the current roadmap for CE? I saw there’s a new CE environment in Win10 IoT, but it feels like too-little-too-late to save the myriad of embedded enterprise CE systems.

There is a recent post about it.

https://techcommunity.microsoft.com/t5/internet-of-things/mo...

> What’s the current roadmap for CE?

No clue sorry! I left many many years ago.

CE7 was incredibly good, the minimum image size we had it running on was ~20MB on disk.

I saw the WPF-esque XAML demos for CE7 - I was really impressed and also hopeful that it would lead to smooth fluid UIs in embedded devices like car infotainment systems and home thermostats.

Oh how naive I was!

(I assume the push for WPF-like XAML in CE7 was because at the time (2011-2012?) there was still hope of a CE7-based Windows Mobile competing with iOS and Android, and Microsoft realized the mistake of not modernizing CE's UI system at-all between 2002 and 2012? Windows Mobile 6.5 was a hideous eyesore (and all the pretty home-screen functionality was literally just skin-deep).

Palm and handspring used moto 68k before PalmOS 5.0, ARM since then

Are you sure?

Intel makes both CPUs and GPUs.

AMD makes both CPUs and GPUs.

nVidia would make both CPUs and GPUs.

What would be the difference?

Actually, as long as there are a number of arm design licensees, the situation would be way better than it currently is for x86: beyond Intel and AMD there is basically no other player in that field (Via? Transmeta? are they even noticeable?).

The difference is that Nvidia would be a customer and the owner of ARM at the same time. An Nvidia controlled ARM will primarily exist to make Nvidia happy at the expense of all other ARM customers.

nVidia already makes CPUs and GPUs (they make the CPU that goes into the Nintendo Switch, for example). Buying ARM doesn’t change much in that regard, except that they can save on licensing fees and it puts them in a position to disrupt competitors who also produce ARM processors.

I would think not licensing new IP would be the main concern. basically hogging ARM IP completely for themselves and their products. It would take a couple of years for companies and devs to switch to a different architecture, but you can be sure that there are competitors willing to fill that gap.

Why would nVidia disrupt competitors?

They need to pay $30bn to buy ARM which means they want to maximize the return for that purchase, which would be selling more and not disrupt for competitors to look elsewhere.

Why give them the opportunity to do so?

What’s at steak? Why is it too much?

It’s not that they will go out of their way to harm it. The ecosystem will change from being the top focus of arm to becoming more of an afterthought. Nvidia arm could end up taking more of a take it or leave it stance because those customers are no longer the top priority.

> Apple won't be phased; they basically design everything themselves

Their new notebooks are still moving to an ARM-based CPU... Even if they enhanced the ARM design they still have an ARM licence.

Apple licenses the architecture(instruction set), but the chip design is theirs, I think

I would assume they had protected themselves before the SoftBank purchase. Also, it's odd that they refer to the new stuff as Apple Silicon.

>Also, it's odd that they refer to the new stuff as Apple Silicon.

Apple supposedly pays intel more for their CPUs than other OEM to avoid having their sticker on Macs; So I don't see this as odd, it's usual Apple branding tactic and from the looks of it(ARM being under threat of sale) it seems to be a wise decision as an average Apple customer wouldn't mind if it switches the architecture for its silicon.

>Apple supposedly pays intel more for their CPUs than other OEM to avoid having their sticker on Macs;

It would be more accurate to say Apple does not receive the Marketing Discount from Intel.

In reality i seriously doubt those discount dont materialise. Intel could have given the TB3 controller for free as part of the deal. Likely less than the Sticker Discount, but still a decent cost saving.

I don't think Intel stickers are mandetory. Some other OEM don't have them either.

Intel gives you a discount for putting stickers on there though.

Apple did make a big thing of it when they switched to Intel.

> "Apple ... could hard fork tomorrow, or two years ago, and nobody will notice"

Could they really do this? Start making incompatible designs? I would have thought that Apple's ARM architecture license would include terms to preclude incompatible forks.

>I would have thought that Apple's ARM architecture license would include terms to preclude incompatible forks.

For an ordinary licensee you might think so, but Apple may well have special terms, given that they co-founded ARM

> given that they co-founded ARM

?????

> The company was founded in November 1990 as Advanced RISC Machines Ltd and structured as a joint venture between Acorn Computers, Apple Computer (now Apple Inc.) and VLSI Technology.[17][18][19] The new company intended to further the development of the Acorn RISC Machine processor, which was originally used in the Acorn Archimedes and had been selected by Apple for its Newton project. […] The company was first listed on the London Stock Exchange and NASDAQ in 1998[23] and by February 1999, Apple's shareholding had fallen to 14.8%.[24]

* https://en.wikipedia.org/wiki/Arm_Holdings#Founding

Holy moly, how did I never know this? I've worked on ARM based systems for years and never run across this tidbit. Thanks for sharing!

From the LA Times in 1990

https://www.latimes.com/archives/la-xpm-1990-11-28-fi-4993-s...

Apple went on to use ARM in the Newton and other devices....

> Apple won't be phased

Nit: fazed, not phased.

I mean, they presumably also won't be phased either

Unless we're in the Star Trek universe...

That doesn’t jive with how I’ve been spelling it....

It's jibe :)

I know. That was the idea. It’s one of those mistakes that is like nails on a chalkboard for me.

I was expecting you to make another one, but "nails on a chalkboard" seems to be the correct usage. Am just ever so slightly disappointed.

Meh, 'I could care less'.

...I can barely bring myself to type and submit that...

I’m (half-)British and I consider “could care less” a hideous American corruption. Recently I successfully argued my case with an American friend, not coincidentally a physicist, by analogy to zero-point energy.

Yes so am I (person you're replying to) - fully - and also consider it both hideous and American.

I wouldn't call it a corruption though, because I've never actually heard it in Britain; just American film and television, and even on HN.

Very related David Mitchell rant:

"Dear America..." https://www.youtube.com/watch?v=om7O0MFkmpw

When I hear someone say it, I often respond that "I could care less, but only a little bit" to make the point about that phrase.

Both "I could care less" and "I couldn't care less" make sense. If you emphasize the second word ("I could care less") and phrase it in a tentative tone of voice it conveys the meaning - i.e. it's possible that I could somehow care less, but it's a theoretical possibility only.

FWIW Professional Linguists think it's perfectly fine to say "I could care less" and that people who peeve about it are just pointless peevers who don't understand how language works.

That's not the point: "I could care less" is grammatically correct, but the meaning is not what was intended. If I had a cup and wanted to tell you there's no more beer in the cup, I wouldn't say "there's beer in the cup", even though the sentence "there's beer in the cup" would be accepted by professional linguists.

> That's not the point: "I could care less" is grammatically correct, but the meaning is not what was intended.

The intended meaning is obvious. There's plenty of colloquial phrases which say one thing but mean another: "I could murder him", for example, doesn't literally mean that. That is how language works.

And, if you saying "there's beer in the cup" was understood by everyone as meaning "there's no more beer in the cup", professional linguists and dictionaries would be A-OK with that interpretation.

By analogy to vectors though, 'I could murder him' is directionally correct but magnitudinally wrong. Whereas 'I could care less' (as used) is directionally wrong - no you couldn't, your point is how little/not at all you care, so stop talking about the existence of an amount that you do care!

> And, if you saying "there's beer in the cup" was understood by everyone as meaning "there's no more beer in the cup", professional linguists and dictionaries would be A-OK with that interpretation.

The fact that "I could care less" is now clearly understood is not intentional. It's an adaptation to a mistake. It's similar to how people say "irregardless", which is clearly an error. It's unambiguous, and everyone knows what you mean. It's still a mistake, and using "regardless" is strictly better.

> is not intentional. It's an adaptation to a mistake.

Yes, that's what language is. Will you be complaining about "apron" as an adaptation to a mistake?

https://www.quickanddirtytips.com/education/grammar/how-a-na...

> It's similar to how people say "irregardless", which is clearly an error.

It's an error that's been being made since 1795, though, and people have been understanding it which pretty much qualifies it as being incorporated into the language.

(I should note that I am strongly descriptivist and think that prescriptivists are silly people.)

> using "regardless" is strictly better.

I think "regardless" is better, yes, because it's a lot easier to say but that's just my opinion, not a Gilt Edged Rule Of English.

that's kind of a stretch; I would imagine "I couldn't care less" is vastly preferred, and the people that use "I could care less" are most likely doing so out of ignorance and not an intent to convey the subtle "theoretical possibility"

Actually, thanks for doing this. As a non-native, frequent misspellings are incredibly annoying as it’s difficult to see which one’s correct.

Now I learned not one but two important correct spellings. Thanks!

You got me

"nVidia's history of not playing nicely with others": can you elaborate or mention an example? That aspect has escaped me thus far.

There was the apple gpu fiasco. Apple laptop GPUs were failing at an extraordinary rate in 2010-2011, and Apple were getting harangued in the media about it (which was justified). Apple eventually announced a program to cover replacement of the GPUs, but it was very delayed. I'm 100% convinced the delay was caused by Nvidia refusing to accept responsibility. When the program was released, in order to have the repair completed at no cost, your computer had to fail a very specific graphics diagnostic, and the failure would be linked to your serial number in Apple's repair CRM. At the time, no other apple repair program had this requirement — Don't get me wrong, it makes sense to use diagnostics to verify hardware issues, but they aren't the only tool you can use. I'm convinced Nvidia only would reimburse apple for computers that had a recorded failure of this specific test.

Since 2014, apple has almost exclusively used AMD graphics cards in computers that do have discrete GPUs, and I don't think it's unreasonable to suggest this was motivated with their terrible experience with Nvidia.

(I'm aware I'm not really backing this up with evidence, but I think the publicly available facts about the whole situation support this narrative)

That situation was also following Nvidia spoiling one of Steve Jobs' announcements even further back, so there was already a bit of bad blood.

Good point, I had forgotten about that.

Which one was that?

Apple also refuses to sign the kernel extension necessary for NVidia GPU drivers to run. NVidia had made a GPU driver for High Sierra, but literally a version later Apple locked down the kernel so that only signed drivers could be loaded. Anyone who was using Nvidia GPUs through, say, an external GPU box got screwed.

I had really hoped that the external GPU thing would takeoff, and give us a good option for gaming on Macs. I bought the dev kit, and it was nothing but a headache, which stopped working at High Sierra too, because they stopped supporting TB2. I'm bitter about it. It really seems like we ought to be "there" by now, but it's starting to look like it will just never be a thing.

It was reported that the problem was NVidia switching to lead free solder for their chip to motherboard connections but not also switching the potting material to one that had the same thermal expansion coefficient as the new solder.

That was a general problem when the industry switched to lead free solder, see also the Red Ring of Death on the Xbox 360.

Don't get me wrong, it makes sense to use diagnostics to verify hardware issues, but they aren't the only tool you can use. I'm convinced Nvidia only would reimburse apple for computers that had a recorded failure of this specific test.

In 2010, Apple wasn’t exactly a beleaguered company. It wasn’t like NVidia not reimbursing Apple would break the bank. Apple should have repaired it just to protect its reputation.

Jobs himself was famous for saying to high level employees that at a certain level of authority, you don’t get to make excuses.

I do think Apple should have just undertaken the repairs much, much earlier. I do also think that it was incredibly complex situation, and my guess is Apple was flabbergasted that Nvidia initially wouldn't reimburse them for the failed parts. So, internally at apple, my guess is that it was about the 'principle' of the matter. They also probably recognized that if they started replacing the boards of their own volition without an agreement in place, they would never be able to get any compensation from nvidia at all.

Again, I agree with you, apple should have launched the program much, much earlier.

But, I mean, that's besides the point because this isn't a post about how Apple handled these failing GPUs, it's about Nvidia's inability to work with others.

> I'm 100% convinced the delay was caused by Nvidia refusing to accept responsibility.

To be fair, the issue was mainly caused by Apples form-over-function thermal design.

To best of my knowledge, Nvidia GPUs works just fine in all other laptops.

100% not true, my Dell Latitude at the time also died. There were major class action lawsuits at the time and basically all the manufacturers had to accept liability and fix them.

No. I had a laptop in 2010 that also suffered from this issue, and there were also PC GPUs affected.

Laptops were only affected en masse because they have far more heat-cool thermal cycles than the average desktop computer.

Discrete GPUs break all the time, customers are just used to it. Maybe it’s improving since GTX1xxx/RX4xx era but they were extremely unreliable when Apple was using them/industry was switching to RoHS compliant Pb free solders.

This issue was industry wide. RIP my XPS M1330.

> To best of my knowledge, Nvidia GPUs works just fine in all other laptops.

Nvidia bumpgate says hello.

While is was probably a factor, I believe it's also because NVIDIA refuses to let Apple modify their drivers to push their Metal API.

I recall the GeForce Partner Program raised some controversy, although I couldn't say whether or it was bad or good or why.

There's also the issue with them refusing to open source Linux drivers. Supposedly this is because they throttle workstation GPUs to create a market for higher-end versions while reducing the manufacturing diversity, but I haven't heard a definitive source for this.

> they throttle workstation GPUs

Not sure about this, but they definitely prevent the loading of drivers for GeForce cards if they detect a hypervisor (Quadro cards work). Nvidia claims that it's a bug, and that they're not fixing it only because GeForce cards are not sold to be run in virtual machines. At the very least it's a dubious claim since for a short time there was an arms race as workarounds were figured out and the next version of the driver would detect them...

Of course it's market segmentation by obscurity so it only lasted a few weeks and it's trivial to work around, but still it's quite shady since there's absolutely no ill effect from the work around.

It is absolutely not a bug, Disassembly of NVIDIA drivers shows that its looking for specific KVM signatures, and if you modify those signatures, magically it starts working again...

Removing KVM signature is now the standard solution. But there's still a working driver patch to remove the virtualization check, it's available at here.

* sk1080/nvidia-kvm-patcher: Fixes "Bug" in Nvidia Driver preventing "Unsupported Configurations" from being used on KVM

https://github.com/sk1080/nvidia-kvm-patcher

Of course it's not a bug. :-)

Not doubting that this is the case, but what's the reason for nVidia wanting to restrict use in virtual machines?

"Fuck you, pay me". Nvidia explicitly forbids "datacenter usage" of their drivers outside of specific licensed-for-datacenter product stacks. (The only exception is for crypto mining, for some reason.) This isn't the only product segmentation in their lineup, either. CAD software optimizations are only available for Quadro cards and up, virtual GPUs are only available for GRID cards, etc. The same GPU they'll charge $600 to gamers for will often go for $3000 or more for businesses.

> The only exception is for crypto mining, for some reason.

I think this is because they were outcompeted by AMD in the GPU crypto days, so they didn't want to get in the way of that (nor were they likely to succeed in blocking this use).

Why does Intel cripple ECC and I/O virtualization [0] in high-end consumer desktop systems? Force people to buy Xeon, even if the performance is otherwise equivalent or higher for the use case. Why does Nvidia restrict virtualization? Force people to buy Quadro and Tesla.

> nVidia

FYI, a few years ago, nVidia has officially changed their name to "Nvidia"...

[0] IOMMU is not only a virtualization feature, it's also an important security feature to protect the host from DMA attacks of malicious peripherals (e.g. 1394, ExpressCard, Thunderbolt, USB 4). Fortunately Intel no longer cripples IOMMU (VT-D) since Skylake, but ECC is another story.

I believe it’s all caps.

http://international.download.nvidia.com/partnerforce-us/Bra...

I stand corrected. I also found a clarification from Wikipedia on the situation.

> "From the mid 90s to early-mid 2000s, stylized as nVIDIA with a large italicized lowercase "n" on products. Now officially written as NVIDIA, and stylized in the logo as nVIDIA with the lowercase "n" the same height as the uppercase "VIDIA".

True, but if isn't an acronym (correct me if I'm wrong) I am not going to use all caps. Because it feels like attention grabbing marketing BS to me.

They don't want people using geforce cards in compute servers/clusters. It's artificial market segmentation since the workstation cards are higher margin. They want customers to pay the big bucks since hypervisor setups are basically only used in business and big budget settings.

> I recall the GeForce Partner Program raised some controversy, although I couldn't say whether or it was bad or good or why.

It was about NVidia wanting to make it so their 3rd party GPU partners, (EVGA, Asus etc.), could only sell their popular GPU brands with NVIDIA in them, so for AMD they'd have to come up with something customers are entirely unfamiliar with.

> here's also the issue with them refusing to open source Linux drivers.

It's not just that. AMD didn't open-source their original Linux driver, presumably there could have been some 3rd party licensing issues, but they wrote a new one that is open-source.

NVIDIA doesn't even let others write an open-source driver for them, they make it purposely difficult to reverse-engineer, sign their firmware that they only release with a massive delay and generally refuse to cooperate.

> It's not just that. AMD didn't open-source their original Linux driver, presumably there could have been some 3rd party licensing issues, but they wrote a new one that is open-source.

Even further than this, they created such a healthy environment for it that there are now two competing open source Vulkan drivers, and the third party one (RADV) is usually winning by a bit, and is now directly supported (with staff) by Valve.

NVIDIA can compete based on the superior performance of their products, but that's not enough. They try to attempt lock-in by encouraging software writers to use CUDA and not use OpenCL.

In many projects that support both, OpenCL and CUDA often give similar performance, and in some, OpenCL is faster than CUDA. So CUDA isn't about getting more performance - it's about excluding everything non-NVIDIA.

I suspect that NVIDIA's gameplaying is what led to Apple deciding to not use NVIDIA in any Macs for the last several years.

They don't need to do any lock-in when the choice is between plain old C with printf like debugging, and a set of programming languages able to target PTX, with out of the box support for C, C++, Fortran and several managed languages, and nice graphical debugging tools.

C++ support only came to OpenCL after it was clear to Khronos that it was time to move beyond C only vision of the world, yet almost no one bothered to implement OpenCL 2.x, which is why OpenCL 3.0 is basically OpenCL 1.2 to make all those OEM compliant, and SYCL is now moving into an backend agnostic C++ stack.

So the competition did not move a finger to make OpenCL eco-system beat CUDA tooling, Google did not cared one second to make OpenCL available on Android instead pushed for their own Renderscript C99 dialect, but it is easy to blame NVidia when competition isn't up to their stuff.

Just like Apple, after creating OpenCL and offering it to Khronos, moved away from their own creation after disputes on how OpenCL roadmap should be managed. Ironically, just like with CUDA, Metal Compute happens to be multiple language aware and having a modern API.

The problem with OpenCL, just like OpenGL, is that it relies on vendor support for essentially the full stack. That means if a single vendor doesn't support the feature, it's basically impossible to ship any code that uses that feature to arbitrary users. Nvidia lagging behind in its support for OpenCL 2.x was probably not explicitly stated as "let's sabotage this competing open standard by dragging our feet" but that was the result. AMD and Intel both put together OpenCL 2.x support, so clearly Nvidia could have finished their incomplete support but didn't.

The fundamental failing Khronos made was repeating the GLSL mistake and not specifying a compiled IR that would decouple OpenCL kernel code from program execution. This was clearly known to be a mistake at the time, given ample evidence from CUDA and HLSL that an IR was possible and could greatly improve the developer experience.

CUDA is probably the biggest thing here, also I don't think they open source much.

>... also I don't think they open source much.

You could say that: https://drewdevault.com/2017/10/26/Fuck-you-nvidia.html

This is increasingly becoming not true. See cutlass and the new release of cufftdx.

These are both things that are unlikely to ever be able to run on anything other than nvidia hardware. It's hardly a mark in their favour over abusive lock-in practices.

Well, of course. They're written in cuda. Why would Nvidia try to write it in a language like opencl that even AMD is moving away from, and instead making fft/blas library calls? At least the source code is there if you wanted to port it.

I think it's very clear that I wasn't commenting on what framework they choose to write their projects in, and you're rather deliberately missing my point.

> These are both things that are unlikely to ever be able to run on anything other than nvidia hardware.

I think it's very clear that the framework was important here. Because you were pointing out that they don't open source much, I pointed out that they do, then you nitpicked by saying it's cuda.

I had no idea, so I just searched and found there are 201 repositories in their github:

https://github.com/NVIDIA

that's a pretty meaningless number without further context

The person I replied to said nvidia doesn't contribute much to open source. I pointed at their github repositories. If you care about the context, it is there, and there is no risk of me summarizing it unfairly.

Having some of your code in github isn't contributing to open source.

nVidia has a very long history of not getting along with Linux.

And now, hilariously enough, they have ended up with a lot of their growth coming from Linux-based systems.

To be fair, I recently bought a new Thinkpad, and with Ubuntu 20.04 and Nvidia CUDA just worked, with absolutely no tuning (I didn't even need to set LD_LIBRARY_PATH, which I found pretty odd). So things seem to be improving, because there's money in it for Nvidia.

The Nouveau drivers have a throttled clock speed due to firmware signatures. And the proprietary drivers are complete mess on wayland, where most compositors and other GPU drivers use GBM, but Nvidia did their own thing and used EGLStreams. Maybe someday in the future the situation will improve, but historically, it hasn't been good.

This sentiment comes entirely from their lack of open source drivers for Linux. It became pretty much a meme after a rant that Linus had about them in 2012.

They have made it even more difficult for open-source to be able to reverse engineer their drivers with signed firmware blobs, push their own solution on Wayland, (ELStreams), over everyone else adopting GBM etc. - in short they refuse to cooperate in an official capacity.

Both Intel and AMD do have an official FLOSS Linux driver and regularly work with the wider community.

Google it maybe? Here's one example thread to pull at: why did Apple stop buying nVidia GPUs across their entire hardware line? EGLStreams is a recent pain in the ass. nVidia's infanticide of OpenCL with CUDA being another painful reminder that nVidia doesn't care about your software needs, just their own.

It's really not hard to find a company that nVidia's pissed off with its dealings.

If there was an infanticide of OpenCL, Apple (the original creator of OpenCL) was at least as responsible as Nvidia.

Interestingly NVIDIA uses a RISC-V core in their GPUs.

Why don’t Apple buy ARM?

They don't need too. They were a cofounder, and have a perpetual arch license

ah, this answers a lot of questions for me :)

This would be ten times worse (for everyone except Apple).

Wouldnt there be IP that Apple needs if they go their own way There was something about how qualcomm Bundles ip licensing with hardware purchases

Apple just spent a billion dollars buying Intel's wireless modem business. If there's still a hardware reliance on Qualcomm, it's certainly coming to a quick end.

I was commenting on the parent comment, that Apple doesn’t need ARM with an example of another company that push their products with their IP licenses

Reply on news.ycombinator.com