The Seawolf was kinda the first attempt at a next generation attack sub while we were still figuring out the technology, making it far too expensive. But it led to the Virginia class, which has gone into mainstream production.
In some ways there's a similar situation with the F-22 vs F-35, though those two may have a bit more of a difference on roles and requirements.
>Notably, AMD's closed-source Vulkan driver currently uses a different pipeline compiler, which is the major difference between AMD's open-source and closed-source Vulkan drivers.
The windows driver has 2 paths, the internal compiler, and the same LLVM as in the open source amdvlk release (though there might be things like not-yet-upstreamed changes, experimental new hardware support etc. that differ from the public version, it was fundamentally the same codebase). The same for DX12 (and any other driver that might use their PAL layer). If you want to confirm you can see all the llvm symbols in the driver's amdvlk{32,64}.dll and amdxc{32,64}.dll files. From what I remember, the internal compiler path is just stripped out for the open source amdvlk releases.
I believe the intent was to slowly deprecate the internal closed compiler, and leave it more as a fallback for older hardware, with most new development happening on LLVM. Though my info is a few months out of date now, I'd be surprised if the trajectory changed that quickly.
Either licensing issues (maybe they don't own all parts of the closed source shader compiler) or fears that Nvidia/Intel could find out things about the hardware that AMD wants to keep secret (the fears being Unfounded doesn't make the possibility of them being a reason any less likely). Or alternatively they considered it not worth releasing it (legal review isn't free) because the LLVM back-end was supposed to replace it anyway.
> or fears that Nvidia/Intel could find out things about the hardware that AMD wants to keep secret (the fears being Unfounded doesn't make the possibility of them being a reason any less likely)
When the fears are unfounded the reason isn't "Nvidia/Intel could find out things about the hardware", it's "incompetence rooted in believing something that isn't true". Which is an entirely different thing because in one case they would have a proper dilemma and in the other they would need only extricate their cranium from their rectum.
> When the fears are unfounded the reason isn't "Nvidia/Intel could find out things about the hardware"
Good luck trying to explain that to Legal. The problem at the core with everything FOSS is the patent and patent licensing minefield. Hardware patents are already risky enough to get torched by some "submarine patent" troll, the US adds software patents to that mix. And even if you think you got all the licenses you need, it might be the case that the licensing terms ban you from developing FOSS drivers/software implementing the patent, or that you got a situation like the HDMI2/HDCP situation where the DRM <insert derogatory term here> insist on keeping their shit secret, or you got regulatory requirements on RF emissions.
And unless you got backing from someone very high up the chain, Corporate Legal will default to denying your request for FOSS work if there is even a slight chance it might pose a legal risk for the company.
> Hardware patents are already risky enough to get torched by some "submarine patent" troll, the US adds software patents to that mix.
Software patents are indeed a scourge, but not publishing source code doesn't get you out of it. Patent trolls file overly broad patents or submarine patents on things they get included into standards so that everyone is infringing their patent because the patent covers the abstract shape of every solution to that problem rather than any specific one, or covers the specific one required by the standard. They can still prove that using binary software because your device is still observably doing the thing covered by the patent.
Meanwhile arguing that this makes it harder for them to figure out that you're infringing their patent actually cuts the other way, because if plaintiffs are clever they're going to use that exact reasoning to argue for willful infringement -- that concealing the source code is evidence that you know you're infringing and trying to hide it.
> And even if you think you got all the licenses you need, it might be the case that the licensing terms ban you from developing FOSS drivers/software implementing the patent, or that you got a situation like the HDMI2/HDCP situation where the DRM <insert derogatory term here> insist on keeping their shit secret, or you got regulatory requirements on RF emissions.
To my knowledge there is no actual requirement that you not publish the source code for radio devices, only some language about not giving the user the option to exceed regulatory limits. But if that can be done through software then it could also be done by patching the binary or using a binary meant for another region, so it's not clear how publishing the code would change that one way or the other. More relevantly, it's pretty uncommon for a GPU to have a radio transceiver in it anyway, isn't it? On top of that, this would only be relevant to begin with for firmware and not drivers.
And the recommended way of implementing DRM is to not, but supposing that you're going to do it anyway, that would only apply to the DRM code and not all the rest of it. A GPU is basically a separate processor running its own OS which is separated into various libraries and programs. The DRM code is code that shouldn't even be running unless you're currently decoding DRM'd media and could be its own optional tiny little blob even if the other 98% of the code is published.
The open source release of amdvlk has never been buildable for windows as all the required Microsoft integration stuff has to be stripped out before release.
So at best it'll be of limited utility for a reference, I can see why they might decide that's just not worth the engineering time of maintaining and verifying their cleaning-for-open-source-release process (as the MS stuff wasn't the only thing "stripped" from the internal source either).
I assume the llvm work will continue to be open, as it's used in other open stacks like rocm and mesa.
I see, but I still don't get why any of that had to be stripped, aren't they using public APIs? Nothing there has to be in any particularly way secret.
Huh, interesting. Very MS style of problem. Why can't they develop public OS interfaces for that? I.e. you don't need the full code for that, just interface libraries. Not having public interfaces in this day and age is simply dumb.
Bluntly: because they don't get software and never did. The hardware is actually pretty good but the software has always been terrible and it is a serious problem because NV sure could use some real competition.
I wish hardware vendors would just stop trying to write software. The vast majority of them are terrible at it and even within the tiny minority that can ship something that doesn't non-deterministically implode during normal operation, the vast majority of those are a hostile lock-in play.
Hardware vendors: Stop writing software. Instead write and publish hardware documentation sufficient for others to write the code. If you want to publish a reference implementation that's fine, but your assumption should be that its primary purpose is as a form of documentation for the people who are going to make a better one. Focus on making good hardware with good documentation.
Intel had great success for many years by doing that well and have recently stumbled not because the strategy doesn't work but because they stopped fulfilling the "make good hardware" part of it relative to TSMC.
The documentation can be published in advance of the product launch.
Intel and AMD did this in the past for their CPUs and accompanying chipsets, when any instruction set extensions or I/O chipset specifications were published some years in advance, giving time to the software developers to update their programs.
Intel still somewhat does it for CPUs, but for GPUs their documentation is delayed a lot in comparison with the product launch.
AMD now has significant delays in publishing the features actually supported by their new CPUs, even longer than for their new GPUs.
In order to have hardware that works on day one, most companies still have to provide specifications for their hardware products to various companies that must design parts of the hardware or software that are required for a complete system that works.
The difference between now and how this was done a few decades ago, is that then the advance specifications were public, which was excellent for competition, even if that meant that there were frequently delays between the launch of a product and the existence of complete systems that worked with it.
Now, these advance specifications are given under NDA to a select group of very big companies, which design companion products. This ensures that now it is extremely difficult for any new company to compete with the incumbents, because they would never obtain access to product documentation before the official product launch, and frequently not even after that.
The problem is, making hardware is hard. Screw something up, in the best case you can fix it in ucode, if you're not that lucky you can get away with a new stepping, but in the worst case you have to do a recall and not just deal with your own wasted effort, but also the wasted downstream efforts and rework costs.
So a lot of the complexity of what the hardware is doing gets relegated to firmware as that is easier to patch and, especially relevant for wifi hardware before the specs get finalized, extend/adapt later on.
The problem with that, in turn, is patents and trade secrets. What used to be hideable in the ASIC masks now is computer code that's more or less trivially disassemblable or to reverse engineer (see e.g. nouveau for older NVDA cards and Alyssa's work on Apple), and if you want true FOSS support, you sometimes can't fulfill other requirements at the same time (see the drama surrounding HDMI2/HDCP support for AMD on Linux).
And for anything RF you get the FCC that's going to throw rocks around on top of that. Since a few years, the unique combination of RF devices (wifi, bt, 4G/5G), antenna and OS side driver has to be certified. That's why you get Lenovo devices refusing to boot when you have a non-Lenovo USB network adapter attached at boot time or when you swap the Sierra Wireless modem with an identical modem from a Dell (that only has a different VID/PID), or why you need old, long outdated Lenovo/Dell/HP/... drivers for RF devices and the "official" manufacturer ones will not work without patching.
I would love a world in which everyone in the ecosystem were forced to provide interface documentation, datasheets, errata and ucode/firmware blobs with source for all their devices, but unfortunately, DRM, anti-cheat, anti-fraud and overeager RF regulatory authorities have a lot of influence over lawmakers, way more than FOSS advocates.
Linux already maintains quirks code paths for all kinds of devices where the manufacturer "could have" updated the firmware to fix the bug but never did.
For the now-fashionable LLMs-on-GPUs world, it's pretty much just matrix multiplications. How many patents can reside in that? I don't expect Google to sell TPUs because that's not the business they're in, but AMD could put them in their SoCs without writing drivers: https://cloud.google.com/tpu/docs/system-architecture-tpu-vm...
Perhaps, but the percentage of Americans foreign-born is at a 100 year high. And the percentage of under 18s who have a foreign-born parent is at an all-time high (25.6% of students, the previous peak was 21.6% in 1920).
And if their children are underperforming in schools it would be important to know.
Well, they're killing the animals before feeding them to the lions etc. So it won't have much of a hunting experience. But perhaps a scavenging experience is enough for them.
Yeah that’s what I meant by it being partial for compassion reasons. They still get the feel of eating the whole body as the article states. But not the feeling of hunting it.
IBM has kept researching in Albany. They license manufacturing technology to other parties even if they gave up trying to build their own fabrication facilities a decade ago.
Ford did try before. For example, the 2006 Ford GT. In comparison, the similarly-priced Ferrari F430 of the era was the mass market automobile with about 5x as many made.
So I wonder how much one should draw from these rationalizations of prestige marketing. At the end of the day Ferrari still uses its larger super-car production capacity to keep even Ford only an occasional entrant into the low-end of market.
Building a halo product that attempts to compete in a luxury niche is very different from running your (whole) business like a luxury niche leader.
We see similar cases with the highest end Chevrolet Corvette. If you look only at statistics and performance numbers, they are arguably competitive, but cheaper. But GM intends to sell as many Corvettes as they can. They do not particularly restrict them in quantity or with a particularly inflated price that represents their role as a status signal or unobtainable luxury piece.
That's why I choose two cars that are at effectively in the same target market at around the same price from the same time period, from Ford and Ferrari.
You're choosing cars (single products) but claiming that businesses are operating in a similar manner. All Ferrari cars are part of how Ferrari operates, while the GT was an exception to how Ford operates.
And I don't dispute that. Simply that Ferrari is the Ford of super cars, building in large quantity for that market.
And I really doubt the applicability of the luxury approach to software. People will vastly overpay to have a 5% faster automobile because it's a personally-owned status symbol and has racing-pedigree. I don't recommend building a company around a new 5% faster database since it isn't a source of prestige for anyone. For many potential customers it'd barely put a mark in their bottom line. And it might be negative if it compromised in any other aspect to obtain it. Benchmarking isn't a reliable winner-take-all marketing approach like racing-based marketing (and it sure isn't as popular).
That still doesn't prove anything. The market demand for the Ford GT could have been 1/2 what Ford produced but the market demand for the F430 was double what Ferrari actually produced.
I am not saying that is actually the case, but production #s just don't really tell the whole story.
Ford only dabbles in these kinds of cars, they have different reasons for picking the # of units they build. They are a lot like Honda, they are not dedicated to this and they instead just build a small # of sports cars every once in a while to try and improve their reputation.
The article glosses over the core founding purpose of each company too. Ford was founded to build massive #s of assembly line built cars at a price the average worker could afford. Ferrari very reluctantly built street cars to fund racing and decided the sport/luxury cars were the best way to make the maximum money with the smallest # of street cars. Less street cars meant less distraction from building the racing cars.
I'm not saying it proves anything but Ferrari is one of the higher volume manufacturers within super cars This applies even compared to many of their competitors we think of as mass market car producers (such as Ford, Toyota, Honda) when they make cars that compete with Ferrari.
As you say that's almost entirely due to demand. But the focus should be on how does Ferrari consistently keep such a large portion of the high end car market? Because that's the key to their success. Many companies can make cars that they can sell out in small volumes at Ferrari prices. But they don't see it scaling, while Ferrari does.
> But the focus should be on how does Ferrari consistently keep such a large portion of the high end car market?
Formula 1
There is no other brands that has as much history in Formula 1 than Ferrari. McLaren is the second one in current constructors at 988 race entries but only started fairly recently (compared to Ferrari) to sell street cars, Lotus a distant 3rd one but hasn't been in formula 1 for the last 30years. Mercedes involvment in Formula 1 is fairly recent. In a way, FERRARI IS FORMULA 1 and FORMULA 1 IS FERRARI. Both need each other. Among other F1 teams and appart from Ferrari, McLaren and Mercedes, only Alpine Renault is fully involved and is definitely not in the luxury sports car market. In the case of Aston Martin it is really much more sponsoring/rebranding of a team than a real constructor as the Aston Martin team, formerly Racing Point, is running a Mercedes Engine. Appart from McLaren which is less known in the street there is simply no other brand like Ferrari in Formula 1. Here are some numbers according to Wikipedia:
_"Ferrari holds the record for the most Constructors' and Drivers' Championships won with sixteen and fifteen, respectively.[8][9] Ferrari also holds the record for the most wins by a constructor with 248,[10] the most pole positions with 253,[11] the most points with 10572,[12] and the most podiums with 834.[13] Ferrari has also entered more Grands Prix than any other constructor with 1114 entries and also maintains the record for the most Grand Prix starts with 1111.[14]"_
Alfa Roméo and Maserati used to be in the same if not better place than Ferrari in the early 50's. And their place in the luxury market constantly faded once they scaled down (being engine suppliers) then stopped their involvment in Grand Prix racing.
Yes, although "larger" here might not be understood by many people. The 360, which was said to be a "larger" production car (by Ferrari standards), only approximately 17,000 of those were ever made.
Making a high performance, expensive, rare car is a small part of the recipe. Most Ferrari owners own more than one Ferrari (65% according to Ferrari of North America study from a decade ago or so, not sure if there's an updated number). A significant number of owners own more than 2. Ferrari famously has dealer prioritization and waitlists that reward you for buying more cars, just like Rolex which another poster mentioned below. When I was servicing these cars around 2015, many customers would buy FFs and Californias that they didn't necessarily want just to have the option to buy something like a first year 458 Spider, not even one of the particularly rare offerings. If someone got something truly low production like a LaFerrari or F12 TDF, chances are that customer had already bought 10+ cars from the dealer and immediately trades in many of their cars at a multi hundred thousand dollar loss to them as soon as they are able to get out of a "regular" model into a special edition. Ferrari drives/meets also serve as social clubs/networking which certainly does provide some positive value for many owners of these vehicles. Very few customers I ever met really cared about outright performance/$ and would have been cross shopping a Ford GT with a 430.
Also they are very specific owners. There's one guy here in LA, specifically the San Gabriel Valley neighborhood who realized there were a ton of visitors from mainland China flush with cash who were either coming as tourists or buying property, i.e. the "fuerdai" phenomenon. Cue a few strategically placed watch/jewelry and lambo/ferrari dealerships later, and I think he's one of the top sellers of their cars in North America...
Man, what I wouldn't give for one of those Ford GTs though. Really cool car imo. The fact that it comes from Ford kinda makes it more exciting to me. Like yeah, a Ferrari is a Ferrari, but wtf is a Ford doing here?? And why does it fit in so well??
AMD is making money on x86 servers.
Is anyone making money doing commodity (i.e. not captive) ARM / R5 server chips for sale? Ampere is losing money and hoping for acquisition.
They’re not hoping anymore. They already got acquired by SoftBank (well technically the deal hasn’t closed yet but I don’t think anyone doubts it’ll close.)
Planned: 29 Completed: 3 Canceled: 26 Active: 3
Yet, unlike the Zumwalts, they are considered a good boat.
reply