FPGAs are mostly used for small series and custom hardware. If you add the software price to the FPGA unit price, it becomes more expensive if you buy, lets say, 1000 FPGAs to ship in your product. Even though you might only license the software once.
It is a trade off, and I have no idea about the state and quality of Vivado. Back in the day I was tinkering with FPGAs, the Xilinx software stack was horrible.
> “Until now, it has been available for free on both Windows and Linux”
If it’s any consolation, it wasn’t and still isn’t available on macOS. Also the part about Linux having a “small user base” made me chuckle.
That’s the opposite of what I’m observing. If they wanted to save costs, they would have dropped Linux support altogether. But instead, they are making it a paid benefit. It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux. Still, there are much better ways to handle this without alienating your user base.
There’s a lot of misinformation in this thread from people who didn’t know about Vivado until this controversy.
> But instead, they are making it a paid benefit. It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux.
Vivado has always been primarily a paid product, including on Linux. The free tier has been a limited version useful for small projects or as a trial, which can support a limited set of FPGAs. The paid tier licenses have price tags in the thousands.
They aren’t making new paid products for a growing user base. They are continuing to support their paid Linux user base.
> It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux.
Well, more correctly that they think the commercial base has grown, and that there's revenue on the table by forcing their standard-edition-using commercial Linux users into contracts.
Maybe the thinking is that the Linux users are more sophisticated and able to self-support than windows shops, and so they're choosing not to buy support even though they could? Seems not implausible, though hard to measure even from within AMD.
Basically this seems like a "good beancounting but terrible marketing" decision out of product management. They're not being deliberately mean to their amateur users, they're just trying to squeeze out a few more dollars for their department's quarterly.
What is really interesting about Linux users is that they cost an enormous amount in support.
I think it was a dev of the reboot of Planetary Annihilation that said their Linux users / build made up a few percent of the sales but over 90 percent of all support tickets (!). Mind you that this was before Valve's Proton.
When I worked at a Linux distro I worked with one device maker who told me in confidence that 90% of their revenue came from Linux-based shops and they only needed a one-person support team. They had a 20 person support team for the remaining 10% of revenue coming from Windows-based shops.
Where I work now the top 10 customers are Linux shops. They probably account for 80% of our revenue. The remaining hundreds of customers are more evenly split between Linux and Windows.
So I guess it depends very much on what industry you're in. For consumer games it might be Windows, but outside of financials and administrative realms and into the world of embedded it's a heckuva lot of Linux. Support costs tend to be lower, and you really only have to target Red Hat and Ubuntu.
If those bugs are only present in the Linux port, then yeah, Linux users cost more to support. But if a significant amount of these bugs affect all platforms, then you could argue that a Linux user is much more valuable to them than a non-Linux user because they provide better feedback. Assuming they actually care about fixing their product.
yes, this. i don't know if it was the same game another another where the devs said that while linux users send the most bug reports they are also the most grateful about having a game that runs on linux, and all their reports genuinely helped make the game better for everyone. (wildly paraphrased from faint memory)
in other words, if you want your game tested and get good feedback for it, do release on linux. maybe even release on linux first. linux users will love you for it, and you get to release a more polished game for a wider audience on windows.
Personally, I think it's probably best to test with Proton as part of a game development cycle to reduce the overall complexity in terms of development (not that games aren't already exceedingly complex). That's just my take... especially if you want to take advantage of that extra 5% or so potential extra market share.
Part of the problem is that Linux isn't really one platform, it's 10 different ones of varying popularity (e.g. supporting Gnome on Debian with Wayland doesn't mean that KDE on Nix with X will work). And it costs somewhere in that 1-10x range to support it because of that.
Yes, and I think a free-version user might produce more support requests than a commercial user for two reasons:
1) commercial/professional users might feel more entitled to support, but typically have a better understanding of linux and more versed in fixing stuff themselves.
-- and more importantly --
2) They probably have a dedicated setup where they can run the AMD-blessed distro
Exactly why we zero asic is making Platypus devices open bitstream and all tooling foss from day one...to protect the world against future evil/dumb version of ourselves.
Of course we don't have silicon yet...so nobody here cares. I think a lot of people forget that Xilinx spent $10B+ develop their awesome devices. I figure we can do it with 1/10th of that.;-)
IDK where it's at now, but 15 years ago xilinx was some of the most garbage software I'd ever worked with. Super buggy, constantly corrupting itself, and this was for me just doing university level projects.
God speed if you can get something a lot better for a lot cheaper.
All the FPGA vendors' tools are pretty bad, and have little incentive to improve because their software is the only option for using their chips, outside of a few niche (and generally quite small) devices.
Quartus was not much better on Linux. Honestly I was really into FPGA's around 2010-12 but gave up as I could not afford the full suites at the time and the software was fucking miserable to work with. They were prone to license amnesia and lord help you after running updates, something likely breaks. OR maybe one day it decides to not work anymore and crash continually. Then you spend hours gnashing your teeth, fighting the install, searching through forums and screaming into the void for help. It was mentally draining.
FPGA software gave me FPGA PTSD. I still to this day don't want to go near them - but I am dying to get back into using them. Help ...
All "software for hardware development" (I'm trying to figure out the right way to describe it) used to be exactly like that. I understand most of it still is.
is there any hope for someone to become a professional in the field without an electronic/computer engineering degree? I'd love to do stuff that I can actually hold in my hand (did some reverse engineering for a usb sound card and it felt rewarding cause I could quite literally hear the success), but I only have a computer science adjacent degree (BBa in Business Information Technology).
You can learn enough CAD to design your first thing in a day, enough about electronics to design your first thing in a week. A 3D printer and an order to jlcpcb.com will get you started. There are a very large number of avenues to go down but those are a couple.
Given how many idiotic ideas are ‘patched’ or worked around in software, it’s probably pretty easy. Especially in the world of modern GPUs when only a handful of people at the factory are able to write drivers for it.
Hardware vendors lost the plot in the Winmodem era.
Winmodems are a very sensible idea. You know software-defined radio? They are that, but for modems. Expandable to support any current or future protocol, at the cost of CPU time. Why do we like SDRs but hate SDMs? That's an irrational position.
The actual problem with winmodems was them breaking the established software/hardware boundary, and the Linux community not having the resources to follow suit.
Nothing stops someone from taking the free Windows Vivado and making it run on Linux, or taking a Winmodem driver and making it run on Linux, or writing a from-scratch software implementation of a 56k modem that can run on any sound card plugged into a phone line (which is what a Winmodem is), or reverse engineering the bitstream format for these FPGAs and writing a compiler from scratch (or even just the device-dependent backend - the frontend and middle-end can be developed in a more normal way and can be shared with other toolchains). But nobody actually stepped up and did it, which I think is proof that the free software community is a lot weaker than it thinks it is.
You could even do it right now, if you wanted to. You're not, and I'm sure there are good reasons for that. Extrapolate it across all developers, and it's unfortunate that it seems none of them have enough reason to do it. On the flip side, if anyone reading this does suddenly decide they have enough reason to do it... (Incentive: FPGAs are fun to play with!)
> Nothing stops someone from taking the free Windows Vivado and making it run on Linux
The EULA and the fact that the linux versior runs faster & has fewer bugs.
> just the device-dependent backend would be a major improvement and the frontend and optimizer could be shared with other toolchains
That's yosys and it's used by smaller commercial vendors.
> or reverse engineering then bitstream format for these FPGAs
Getting the timing is the hard part (+ good routing afterwards). The bitstream format has AFAIK mostly been reversed. 7 series has mediocre support , but US, US+ and Versal doesn't (probably because they're too expensive for personal usage).
It's a practical position. The era of Winmodems was a long time ago, and the hardware was terrible compared to what we have now. Today, SDR is a fun thing anybody can pick up as a hobby. If it doesn't work, I have three other ways of accessing the Internet. If your winmodem didn't work, you didn't have a smartphone or a tablet to connect to ChatGPT (well, Altavista at the time) with and look for help. Then, when they did work, they were really bad because the single core CPUs of the day didn't have multiple cores to have the CPU cycles to run the software end of the modem and do anything else at the time. Which meant if you were running a game (Tuxracer, perhaps? Linmodem support wasn't broad, but it existed) at the same time, you lost. That tends to cause people to not like the product.
I imagine it's due to having had decent enough GPUs and decent enough CPUs, from a single vendor.
If you want the platform to be x86 but not AMD then your only other choice is Intel, but they've only recently started making high performance GPUs. So then you need another vendor for the GPU, and your only choice is Nvidia.
A lot simpler, cheaper and predictable to go with a single vendor for both I imagine?
AMD also had the strongest offering for GPU and CPU using the same memory with the same address space. That allows you to switch between CPU and GPU processing for the same data, without paying the cost of moving the data to and from the GPU. Similar to what we now have on Apple silicon
They tried to push the same into the desktop market with their APUs, where it was mostly ignored. But console games only target a couple hardware configurations, making it viable to take advantage of such hardware features
Also also, AMD’s play has always been to produce HW that offers good performance/$, with the downside of having much weaker SW offerings to go with it.
Consoles are always pressured to minimize upfront purchase costs, and they generally replace the vendor-provider SW stack with their own anyways.
And they’ve been in a rough spot at various times in the past, which probably made them willing to negotiate with the console companies.
Actually looking at this thread, there’s a lot of good reasons they were the go-tos for consoles. Consoles seem to be in rough shape at the moment, I wonder if part of that is that AMD has been doing too well since Zen, haha.
You’re approaching this as if every company had the same corporate intentions.
Nvidia never cared much for those types of deals. They preferred to lose Apple as a business than to admit fault, they’ve always refused to compete on price for the business of Sony and Microsoft’s consoles. They’re adamant to beat at the sound of their own drum.
Non-paying users aren’t customers, though, so they must view all this outrage seems irrelevant. Which suggests that they view free-tier Linux users as significantly less likely to ever pay for its use. That matches my understanding of the (non-Steam) Linux as a miserly and demanding target market, so I don’t really fault them for the choice — especially given how brutally expensive it is to support the IDIC of Miscellaneous Linuxes. Kind of surprised they haven’t just withdrawn free support for anything but Steam Linux, in order to lower their costs (and to produce a ‘free’ build that anyone can run privately but doesn’t interop at all with enterprise). Maybe they want it to be a ‘shareware trial’ for enterprise? Or perhaps they just haven’t thought of it yet.
The "free" version of Vivado is used to develop for Xilinx/AMD's lower tier FPGAs. While offering what I assume are lower profit margins, these lower tier FPGAs make up a large portion of Xilinx/AMD's chip volume.
Xilinx/AMD charging for any of their tools is also a recent thing. 20 years ago, you could download these tools freely without even having to register on their website.
Hardware isn’t where the margins are, and probably is somewhat of a loss leader for small-batch users; for hobby users I would hazard a guess that they’re running at around -10% profit on small sales to try and drive subscription revenue multipliers, and for already-paying users this change is essentially irrelevant and will have zero downside impact on sub revenue. Terrible way to run a profitable business if you fuck up the hardware undercut, but if you can get away with it, subscriptions are certainly a valid answer to maintenance of the platform over time. I still think they didn’t go far enough to make a meaningful dent in conversions from free to paying, though.
(Note that mention of Steam Linux is not about the games aspect, but about the Valve’s seeming plans to become a competitive target for Linux support to the exclusion of other consumer-focused miscellaneity. But I tend to go on about this too often, and shouldn’t have invoked it here, apologies.)
This is not the case here as the software is required to use the hardware they’re selling at any quantity. The software is cost entirely for them, if you’re not buying the hardware you’re not using the software. Given they support Linux for the paid version, its development is already paid for. Absolutely say you won’t provide support for free tier users. Today’s free tier users are tomorrow’s purchasing managers. FPGA is not a big market, so you have to capture comparatively few people for each unit of market share. Good silicon without good software is just very expensive sand.
I doubt that. Dev boards are often not very high margin despite their costs, but absolutely the majority of their profit is from hardware sales, not software licenses. Small volume customers are a combination of a long tail and a loss leader for a marketing pipeline, and FPGAs are almost by definition something where you can't ignore that part of the market (because you only use an FPGA if you can't use anything else and you don't have the volume/margin to justify an ASIC, so it's a niche of niches).
These chips scale up into the price range of $100,000 per chip, I'm not kidding. You really want someone to pick your $100k chip instead of Intel's $100k chip. A single high-end deal offsets the entire tool sales over the entire lifetime of Vivado.
It might be excusable that they want to vet their customers receiving the tool chain for the high end chips to avoid leaking trade secrets to Intel, but that isn't excusable for the low end. Someone who starts with your $10 chips is likely to develop brand loyalty and if they need $100k chips later, they'll be more likely to pick your ones.
Lol, AMD FPGA's the same chip can cost you from $2000 to $10000 depending on which channel you buy it from, or what relationship you have with AMD. Don't be ridiculous. This is an extreme margin business. And especially small time buyers are completely skinned and trimmed of all fat here. That's why many hobbyists are using unsoldered/re-used chips from accelerators or discarded smart network cards, or other second-hand whatever, repackaged in china by smart hands.
> probably is somewhat of a loss leader for small-batch users
Wrong. AMD/Xilinx doesn't sell devices directly to customers, they sell them to distributors in huge quantities. Those distributors then sell them to "small-batch users", and they're not involved with AMD/Xilinx free-tier software at all.
> they’re running at around -10% profit on small sales to try and drive subscription revenue multipliers
Your elided quote removes the five words where I declared my views as speculation openly and in plain language. The complete sentence that you misquoted opens with that:
> I would hazard a guess
I’m perfectly content to be wrong at HN; it’s a forum where we all have opinions and people rarely restrict themselves to exclusively their own expert subjects, or else we’d all never learn anything! So I will be considering the arguments made here by others before engaging with this topic in the future.
Edit: https://news.ycombinator.com/item?id=48256417 from a few days ago apparently makes some of my points much more clearly, or at least with less hostile replies. I wish I’d found it sooner, but I didn’t realize this entire post was a dupe in time to go back through its comments in detail. Would have saved me commenting at all! Ah well.
Whelp, I’m an embedded engineering consultant and will no longer recommend these products to my customers. Or rather, I will ask them to avoid these products entirely.
AMD, you can make more money selling chips than software, but take away the entry level software and you eliminate the on-ramp. I’m not buying a license to prototype.
Very few, entreprise users (aka volume) will pay the license, hobbyists will pirate it if need be.
AMD doesn't want to do support for the hobbyists for free, that's all.
I mean, over the years, I have purchased (or advised other people to purchase) multiple Nvidia GPUs for compute workloads.
And the reason pretty much always came down to good integration between various open source software and proprietary CUDA drivers. And the assumption is that this support will continue for many years.
So, yeah, burning their existing FPGA users is a strong signal never to invest real money in their GPUs for compute workloads.
While Vivado/Vitis etc do amazing things, I challenge anyone to find a person who enjoys using them without TCL interfaces.
These tools do need attention, it's too bad there's not a better model than subscription bases like these.
Pretty sure, based on TCL base, that these tools were native Unix at some point, so the no-linux-free-beer vs windows-free-beer version are hilarious...
Ultimately one has, with so many vendor tools, a windows box somewhere so make it a remote compile machine.
As far as I can tell, Lattice products are the best if you're interested in using F/OSS tools. Yosys / Icestorm supports the Lattice ICE40 and ECP5 devices. QuickLogic also apparently do some officially supported F/OSS stuff for their devices, but I don't know much about them. Think I might have ordered one of their dev boards a while back, and then promptly put it on a shelf and still haven't gotten around to touching it. :-(
The problem is, none of these devices can match the raw performance of the devices from Xilinx / Altera / Microchip / etc. So you're kinda limited if you are trying to do something that needs serious horsepower.
Folks feel outrage when companies start charging for things that were once free.
Okay, but what if you run a company whose business model no longer supports giving away free stuff? How can you transition? What would users consider less outrageous?
AMD isn't giving away free stuff. They are selling FPGA hardware. But further, the free stuff has a lot of restrictions around it which practically gear it only for university usage. And the reason they do that is they want to have new graduates have experience with their software so they can demand it from their future employers.
Basic Vivado is the bare minimum to develop for their hardware. A large amount of functionality is still locked away behind paid IP.
Most of the revenue comes from the IP cores.
A common business model for companies like this is to enable developers to learn their tools cheaply, so that when they develop something for their employer, they're more likely to reach for those tools/ecosystems and have the employer pay for those tools.
This just cuts out beginner/hobbyist FPGA devs from using industry standard tooling.
You need to buy their hardware to use it. In fact there is no way to use most of their hardware without Vivado. So it's more like they are blocking you from using things you've already paid for
The software is useless without their chips and the chips cost a fortune. It's not "business model no longer supports giving away free stuff". It's just bean-counters cutting corners.
When AMD bought Xilinx I was hoping they'd open up the software side like they (eventually) did with their GPU drivers. Looks like that isn't happening anytime soon.
It seems silly to put up SW barriers for people to use your fairly expensive HW, but what do I know.
But you just showed you have deep pockets and they think they can get you to open it again every year for the rest of time.
Xilinx was never positioned that it made sense for them to open it up. If/when it gets run into the ground by AMD short sightedness they might just open it to claim that was the plan all along...
Pretty sure this 'article' was written by an LLM, having scraped the HN discussion on here from 4 days ago. Nothing new there apart from a clickbait title and a ton of ads.
NVIDIA ended support for their 10xx series [1]. To be clear, AMD also moved support for their equivalent 5xxx series to legacy drivers [2], but "supports their cards for many years" doesn't hold value if both companies stopped their respective GPUs at basically the same time.
Also remember that one of those 2 companies has opensource drivers for Linux for their old GPUs, while the other doesn't (newer NVIDIA GPUs have an opensource driver but this isn't the case for the 10xx series). Users of legacy NVIDIA cards needs on Linux needs to use their old driver branches, with results that are less than optimal to say the least.
Your "true observation" doesn't contribute to the context of this particular topic thread which "has nothing to do with [your] comment", as you are "well aware". You should review the HN Guidelines: https://news.ycombinator.com/newsguidelines.html
> Its comparatively smaller user base means less commercial pressure, making it an easy target to throw under the bus whenever companies feel like cutting costs or boosting profits.
Eventually the empire will strike back though.
I now marked AMD as a company that can not be trusted.
We need more indie companies in general, and cheap 3d printing for the masses. It'll be a long way to go to nanoscale perfection, but we'll have to go it - AMD but also Intel before, showed that NONE of those mega-greedy corporations can be trusted. They'll always try to do a switcheroo move. But as I stated here: the empire will strike back eventually. The Barbara Streisand effect is real.
I mean perhaps the silver lining is the projects I use are all stuck on 2022.1 for now. I wonder if this is because they want to gate usage by AI agents.
if "the ai" was delivering value you'd think it would be lowering both cost to develop these dev tools as well as the cost to support them. In that light the move is odd unless they don't know how to use AI to lower those costs or it cant.
We shouldn't wait for them to get their act together, as it's in the best interest of a cartel to not have competitors, compatibility, and transparency.
It should be required after a certain amount of time that schematics and code be open sourced and that anti-walled-garden measures are prevalent so we get compatibility and extensibility right out of the box.
They want their *own* CUDA situation. if you're doing FPGA stuff, the Xilinx hardware is good and Vivaldi is really the only way to use it. AFAIK the open-source fpga situation is pretty far behind it (please correct me if I'm wrong, I work in a place that uses Xilinx FPGAs)
What better way to do that than decrease availability, I ask, both rhetorically and sarcastically! CUDA-proper did well, at least partially, because it was put in front of everyone. This is taking an exclusivity angle that doesn't make much sense for ecosystem development, IMO.
I suspect this goes to show how much influence/priority B2B carries, my point is it's a mistake they've made before. The maintenance/support costs I believe they're trying to remove have paid dividends for CUDA and countless others. Anyway, a play on the tagline before I go: "Together we advance... licensing and removal of [positive] feedback loops over product development"
I seriously think the CUDA situation was set up intentionally.
Jensen Huang (Nvidia CEO) is related to Lisa Su (AMD CEO).
A decade ago, I saw a demo of some AMD skunkworks GPU datacenter tech, that could execute CUDA natively on AMD/ATI graphics cards. Initially was half speed, but having the flexibility was crazy amazing. Created a big buzz in the big iron and educational markets.
Where'd it go? Buried. You cant even find articles about it. Its a few comments on edtech datacenters.
Now look at AMD's graphics line. Where's their ROCm LLM tooling? It's a fucking joke. Its like they're intentionally sinking it for her Nvidia uncle Huang. And Su takes the cheaper CPU market and offers better features than Intel.
I'm not trying to dismiss what you're saying as a possibility (AMD's behavior in many regards over the last 15 years or so is baffling to the point that a family conspiracy feels surprisingly plausible) but Huang isn't Su's uncle.
They are "first cousins once removed" meaning that Su is the child of one of Huang's cousins. Or put another way, one of Huang's grandparents is one of Su's great-grandparents.
Honestly, my mind could never wrap around genealogy and family relations well at all. I thought they were indirect cousin/uncle , or as you say, cousins once removed.
Also my understanding of many Asian cultures is they tend to have a much more tight-knit large family structure. And doubling that is the fact they're 2 heads of world-level hardware tech companies.
And, well, there's no such thing as coincidences. Having all of this line up, and for "some reason" AMD keeps missing when they could have owned a big chunk of the market has a certain family oligopoly smell to me.
Well, but there absolutely are such things as coincidences. It is obvious that coincidences are to be expected.
Combinatorial mathematics actually says certain types of coincidence happen more often than we seem to expect intuitively (eg. the Birthday "Paradox").
I have no particular view on AMD. But any argument that includes "I don't believe in coincidences" should probably be weakened in your estimation.
There's no free alternatives, because AMD doesn't document the bitstream format (i.e. what you need to push to the FPGA to program it to do wha you want).
Xilinx has the best silicon. Everyone else is behind. Altera is basically dead thanks Intel. Lattice is nice for low power but performance-wise they are behind. Don't know much about Microchip, but from the little I've heard their tooling is a disaster even by the standards of FPGA tooling. Then there are Gowin (not bad, but Chinglish docs and everything), Gatemate (pretty innovative and vendor-backed nextpnr support - but only one low-mid FPGA with a promise to release chiplet assemblies of it latter). And Effinix - don't know much about them, do anyone have experience?
This software seems to never have been open source/freely licensed. That's not a bait and switch. They were giving you a commercial product, for free, and now have decided not to.
It's likely a case where maintaining separate builds for the free and commercial tiers was getting complex. Often times, this kind of software requires lots of manual reviewing and adding or removing modules, and they probably decided it's just not worth it.
I don't see how that particular line of thinking applies when:
1) They continue to have a free version for Windows
2) They continue to have a version for Linux
I just can't see that cost of having a free Linux version (on top having a paid Linux version) is big?
I literally explained the thinking that the free builds on Linux aren't worth it. If you've ever shipped production software, you'd know this. Just because there's a free build available for Windows doesn't mean it costs the company $0 to release the free build. It's a lot of extra time and QA for each variant of a release. There might be many differences between the Windows and Linux builds, such as the Linux builds require proprietary 3rd party code with royalties, and they chose not to eat that cost.
There's no bait and switch. It's just people expecting things for free, as always, when this was never an open source project.
> It's a lot of extra time and QA for each variant of a release.
In Vivado, it's the same release for the free and expensive builds on both Linux and Windows. It's just a question of the installed license file/license limits.
> such as the Linux builds require proprietary 3rd party code with royalties, and they chose not to eat that cost.
This ignores that their revenue comes from selling chips. The software enables that. No software, no chip sales. This can only have a negative impact on their sales.
Think academic and small companies who don't pay for support opening corner case issues all the time publicly. They want none of the complex support unless you pay (reasonable imo).
And for those who forget RHEL for instance has to pay salaries to back port fixes and such and the same logic applies here.
Xilinx is/was an FPGA company until AMD bought them. Their primary revenue stream is selling chips. This is the equivalent of going back to the days of paid C/C++ compilers (anybody else remember that?).
AMD is not a good company. They stopped innovating after Intel was put down. Except, now Intel has govt backing while AMD will face significantly more competition from not only x86 but arm. Stock price says otherwise but I think they had more than enough time to catch up to Nvidia and simply refused to compete.
AMD business should be selling FPGA hardware. The software suite should come for free. If it doesn't then people should not purchase AMD FPGAs.
It is absurd that in 2026 you have to pay for such tools. It feels like buying a propietary compiler in 80's or 90's.
No one wants that anymore.
FPGAs are mostly used for small series and custom hardware. If you add the software price to the FPGA unit price, it becomes more expensive if you buy, lets say, 1000 FPGAs to ship in your product. Even though you might only license the software once.
It is a trade off, and I have no idea about the state and quality of Vivado. Back in the day I was tinkering with FPGAs, the Xilinx software stack was horrible.
Not to mention, the software is horrendous to use
Bloated and slow, but almost everything is automatable via TCL. A lot of the other profssional SW can only boast about being bloated and slow.
Once you learn how to get automation to work for you, it's not that bad.
Well, at least they don't really try hard to protect their software, unlike some other brands.
> “Until now, it has been available for free on both Windows and Linux”
If it’s any consolation, it wasn’t and still isn’t available on macOS. Also the part about Linux having a “small user base” made me chuckle.
That’s the opposite of what I’m observing. If they wanted to save costs, they would have dropped Linux support altogether. But instead, they are making it a paid benefit. It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux. Still, there are much better ways to handle this without alienating your user base.
There’s a lot of misinformation in this thread from people who didn’t know about Vivado until this controversy.
> But instead, they are making it a paid benefit. It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux.
Vivado has always been primarily a paid product, including on Linux. The free tier has been a limited version useful for small projects or as a trial, which can support a limited set of FPGAs. The paid tier licenses have price tags in the thousands.
They aren’t making new paid products for a growing user base. They are continuing to support their paid Linux user base.
> It can only mean that their Linux user base is growing, ie. more commercial operators are turning to Linux.
Well, more correctly that they think the commercial base has grown, and that there's revenue on the table by forcing their standard-edition-using commercial Linux users into contracts.
Maybe the thinking is that the Linux users are more sophisticated and able to self-support than windows shops, and so they're choosing not to buy support even though they could? Seems not implausible, though hard to measure even from within AMD.
Basically this seems like a "good beancounting but terrible marketing" decision out of product management. They're not being deliberately mean to their amateur users, they're just trying to squeeze out a few more dollars for their department's quarterly.
What is really interesting about Linux users is that they cost an enormous amount in support.
I think it was a dev of the reboot of Planetary Annihilation that said their Linux users / build made up a few percent of the sales but over 90 percent of all support tickets (!). Mind you that this was before Valve's Proton.
Edit: It was <0.1% sales but 20% of all support tickets: https://xcancel.com/bgolus/status/1080213166116597760
When I worked at a Linux distro I worked with one device maker who told me in confidence that 90% of their revenue came from Linux-based shops and they only needed a one-person support team. They had a 20 person support team for the remaining 10% of revenue coming from Windows-based shops.
Where I work now the top 10 customers are Linux shops. They probably account for 80% of our revenue. The remaining hundreds of customers are more evenly split between Linux and Windows.
So I guess it depends very much on what industry you're in. For consumer games it might be Windows, but outside of financials and administrative realms and into the world of embedded it's a heckuva lot of Linux. Support costs tend to be lower, and you really only have to target Red Hat and Ubuntu.
If those bugs are only present in the Linux port, then yeah, Linux users cost more to support. But if a significant amount of these bugs affect all platforms, then you could argue that a Linux user is much more valuable to them than a non-Linux user because they provide better feedback. Assuming they actually care about fixing their product.
yes, this. i don't know if it was the same game another another where the devs said that while linux users send the most bug reports they are also the most grateful about having a game that runs on linux, and all their reports genuinely helped make the game better for everyone. (wildly paraphrased from faint memory)
in other words, if you want your game tested and get good feedback for it, do release on linux. maybe even release on linux first. linux users will love you for it, and you get to release a more polished game for a wider audience on windows.
Personally, I think it's probably best to test with Proton as part of a game development cycle to reduce the overall complexity in terms of development (not that games aren't already exceedingly complex). That's just my take... especially if you want to take advantage of that extra 5% or so potential extra market share.
Part of the problem is that Linux isn't really one platform, it's 10 different ones of varying popularity (e.g. supporting Gnome on Debian with Wayland doesn't mean that KDE on Nix with X will work). And it costs somewhere in that 1-10x range to support it because of that.
Most software picks 1-2 distros (usually LTS Ubuntu and/or Fedora) to support in the default configuration.
That single, annecdotal evidence gets posted everytime a topic like that comes up. I would really love to get more reliable stats about this.
Even the "0.1% of sales" figure is quite atypical.
20% of auto-reported crashes. So it's not like Linux users were writing more tickets but the game crashing on Linux more (because of gfx drivers).
Yes, and I think a free-version user might produce more support requests than a commercial user for two reasons: 1) commercial/professional users might feel more entitled to support, but typically have a better understanding of linux and more versed in fixing stuff themselves. -- and more importantly -- 2) They probably have a dedicated setup where they can run the AMD-blessed distro
Would you rather that the 100% of bugs in those 20% of tickets never got fixed?
Linux users write highly detailed bug reports because it's the only way to get things fixed without coding the fix yourself.
I don't think gaming on Linux is comparable here.
Exactly why we zero asic is making Platypus devices open bitstream and all tooling foss from day one...to protect the world against future evil/dumb version of ourselves.
https://www.zeroasic.com/platypus https://www.zeroasic.com/projects/wildebeest https://www.zeroasic.com/projects/logik
Of course we don't have silicon yet...so nobody here cares. I think a lot of people forget that Xilinx spent $10B+ develop their awesome devices. I figure we can do it with 1/10th of that.;-)
IDK where it's at now, but 15 years ago xilinx was some of the most garbage software I'd ever worked with. Super buggy, constantly corrupting itself, and this was for me just doing university level projects.
God speed if you can get something a lot better for a lot cheaper.
All the FPGA vendors' tools are pretty bad, and have little incentive to improve because their software is the only option for using their chips, outside of a few niche (and generally quite small) devices.
Quartus was not much better on Linux. Honestly I was really into FPGA's around 2010-12 but gave up as I could not afford the full suites at the time and the software was fucking miserable to work with. They were prone to license amnesia and lord help you after running updates, something likely breaks. OR maybe one day it decides to not work anymore and crash continually. Then you spend hours gnashing your teeth, fighting the install, searching through forums and screaming into the void for help. It was mentally draining.
FPGA software gave me FPGA PTSD. I still to this day don't want to go near them - but I am dying to get back into using them. Help ...
All "software for hardware development" (I'm trying to figure out the right way to describe it) used to be exactly like that. I understand most of it still is.
It's long been said:
"AMD never misses a chance to miss a chance."
In this case, the chance to trash its reputation with customers.
especially their marketing dept which made this decision seems to be run by absolute buffoons
I bet this decision was made by showing an Excel sheet.
AMD has long been the proof that hardware is easier than software. Apparently, hardware is also easier than marketing.
I wish more software engineers found out how easy hardware actually is.
is there any hope for someone to become a professional in the field without an electronic/computer engineering degree? I'd love to do stuff that I can actually hold in my hand (did some reverse engineering for a usb sound card and it felt rewarding cause I could quite literally hear the success), but I only have a computer science adjacent degree (BBa in Business Information Technology).
You can learn enough CAD to design your first thing in a day, enough about electronics to design your first thing in a week. A 3D printer and an order to jlcpcb.com will get you started. There are a very large number of avenues to go down but those are a couple.
Given how many idiotic ideas are ‘patched’ or worked around in software, it’s probably pretty easy. Especially in the world of modern GPUs when only a handful of people at the factory are able to write drivers for it.
Hardware vendors lost the plot in the Winmodem era.
Winmodems are a very sensible idea. You know software-defined radio? They are that, but for modems. Expandable to support any current or future protocol, at the cost of CPU time. Why do we like SDRs but hate SDMs? That's an irrational position.
The actual problem with winmodems was them breaking the established software/hardware boundary, and the Linux community not having the resources to follow suit.
Nothing stops someone from taking the free Windows Vivado and making it run on Linux, or taking a Winmodem driver and making it run on Linux, or writing a from-scratch software implementation of a 56k modem that can run on any sound card plugged into a phone line (which is what a Winmodem is), or reverse engineering the bitstream format for these FPGAs and writing a compiler from scratch (or even just the device-dependent backend - the frontend and middle-end can be developed in a more normal way and can be shared with other toolchains). But nobody actually stepped up and did it, which I think is proof that the free software community is a lot weaker than it thinks it is.
You could even do it right now, if you wanted to. You're not, and I'm sure there are good reasons for that. Extrapolate it across all developers, and it's unfortunate that it seems none of them have enough reason to do it. On the flip side, if anyone reading this does suddenly decide they have enough reason to do it... (Incentive: FPGAs are fun to play with!)
> Nothing stops someone from taking the free Windows Vivado and making it run on Linux
The EULA and the fact that the linux versior runs faster & has fewer bugs.
> just the device-dependent backend would be a major improvement and the frontend and optimizer could be shared with other toolchains
That's yosys and it's used by smaller commercial vendors.
> or reverse engineering then bitstream format for these FPGAs
Getting the timing is the hard part (+ good routing afterwards). The bitstream format has AFAIK mostly been reversed. 7 series has mediocre support , but US, US+ and Versal doesn't (probably because they're too expensive for personal usage).
It's a practical position. The era of Winmodems was a long time ago, and the hardware was terrible compared to what we have now. Today, SDR is a fun thing anybody can pick up as a hobby. If it doesn't work, I have three other ways of accessing the Internet. If your winmodem didn't work, you didn't have a smartphone or a tablet to connect to ChatGPT (well, Altavista at the time) with and look for help. Then, when they did work, they were really bad because the single core CPUs of the day didn't have multiple cores to have the CPU cycles to run the software end of the modem and do anything else at the time. Which meant if you were running a game (Tuxracer, perhaps? Linmodem support wasn't broad, but it existed) at the same time, you lost. That tends to cause people to not like the product.
Should be the first of the two chances for the phrase to work.
I'm even surprised they have so much of the console market
I imagine it's due to having had decent enough GPUs and decent enough CPUs, from a single vendor.
If you want the platform to be x86 but not AMD then your only other choice is Intel, but they've only recently started making high performance GPUs. So then you need another vendor for the GPU, and your only choice is Nvidia.
A lot simpler, cheaper and predictable to go with a single vendor for both I imagine?
AMD also had the strongest offering for GPU and CPU using the same memory with the same address space. That allows you to switch between CPU and GPU processing for the same data, without paying the cost of moving the data to and from the GPU. Similar to what we now have on Apple silicon
They tried to push the same into the desktop market with their APUs, where it was mostly ignored. But console games only target a couple hardware configurations, making it viable to take advantage of such hardware features
Also also, AMD’s play has always been to produce HW that offers good performance/$, with the downside of having much weaker SW offerings to go with it.
Consoles are always pressured to minimize upfront purchase costs, and they generally replace the vendor-provider SW stack with their own anyways.
And they’ve been in a rough spot at various times in the past, which probably made them willing to negotiate with the console companies.
Actually looking at this thread, there’s a lot of good reasons they were the go-tos for consoles. Consoles seem to be in rough shape at the moment, I wonder if part of that is that AMD has been doing too well since Zen, haha.
You’re approaching this as if every company had the same corporate intentions.
Nvidia never cared much for those types of deals. They preferred to lose Apple as a business than to admit fault, they’ve always refused to compete on price for the business of Sony and Microsoft’s consoles. They’re adamant to beat at the sound of their own drum.
Nvidia was so thirsty for an x86 license, for years, that it wouldn't consider anything else.
Non-paying users aren’t customers, though, so they must view all this outrage seems irrelevant. Which suggests that they view free-tier Linux users as significantly less likely to ever pay for its use. That matches my understanding of the (non-Steam) Linux as a miserly and demanding target market, so I don’t really fault them for the choice — especially given how brutally expensive it is to support the IDIC of Miscellaneous Linuxes. Kind of surprised they haven’t just withdrawn free support for anything but Steam Linux, in order to lower their costs (and to produce a ‘free’ build that anyone can run privately but doesn’t interop at all with enterprise). Maybe they want it to be a ‘shareware trial’ for enterprise? Or perhaps they just haven’t thought of it yet.
The "free" version of Vivado is used to develop for Xilinx/AMD's lower tier FPGAs. While offering what I assume are lower profit margins, these lower tier FPGAs make up a large portion of Xilinx/AMD's chip volume.
Xilinx/AMD charging for any of their tools is also a recent thing. 20 years ago, you could download these tools freely without even having to register on their website.
Vivado is an IDE for programming AMD FPGAs. One cannot use it without buying AMD hardware.
Hardware isn’t where the margins are, and probably is somewhat of a loss leader for small-batch users; for hobby users I would hazard a guess that they’re running at around -10% profit on small sales to try and drive subscription revenue multipliers, and for already-paying users this change is essentially irrelevant and will have zero downside impact on sub revenue. Terrible way to run a profitable business if you fuck up the hardware undercut, but if you can get away with it, subscriptions are certainly a valid answer to maintenance of the platform over time. I still think they didn’t go far enough to make a meaningful dent in conversions from free to paying, though.
(Note that mention of Steam Linux is not about the games aspect, but about the Valve’s seeming plans to become a competitive target for Linux support to the exclusion of other consumer-focused miscellaneity. But I tend to go on about this too often, and shouldn’t have invoked it here, apologies.)
This is not the case here as the software is required to use the hardware they’re selling at any quantity. The software is cost entirely for them, if you’re not buying the hardware you’re not using the software. Given they support Linux for the paid version, its development is already paid for. Absolutely say you won’t provide support for free tier users. Today’s free tier users are tomorrow’s purchasing managers. FPGA is not a big market, so you have to capture comparatively few people for each unit of market share. Good silicon without good software is just very expensive sand.
I doubt that. Dev boards are often not very high margin despite their costs, but absolutely the majority of their profit is from hardware sales, not software licenses. Small volume customers are a combination of a long tail and a loss leader for a marketing pipeline, and FPGAs are almost by definition something where you can't ignore that part of the market (because you only use an FPGA if you can't use anything else and you don't have the volume/margin to justify an ASIC, so it's a niche of niches).
These chips scale up into the price range of $100,000 per chip, I'm not kidding. You really want someone to pick your $100k chip instead of Intel's $100k chip. A single high-end deal offsets the entire tool sales over the entire lifetime of Vivado.
It might be excusable that they want to vet their customers receiving the tool chain for the high end chips to avoid leaking trade secrets to Intel, but that isn't excusable for the low end. Someone who starts with your $10 chips is likely to develop brand loyalty and if they need $100k chips later, they'll be more likely to pick your ones.
Lol, AMD FPGA's the same chip can cost you from $2000 to $10000 depending on which channel you buy it from, or what relationship you have with AMD. Don't be ridiculous. This is an extreme margin business. And especially small time buyers are completely skinned and trimmed of all fat here. That's why many hobbyists are using unsoldered/re-used chips from accelerators or discarded smart network cards, or other second-hand whatever, repackaged in china by smart hands.
> Hardware isn’t where the margins are
Baseless speculation
> probably is somewhat of a loss leader for small-batch users
Wrong. AMD/Xilinx doesn't sell devices directly to customers, they sell them to distributors in huge quantities. Those distributors then sell them to "small-batch users", and they're not involved with AMD/Xilinx free-tier software at all.
> they’re running at around -10% profit on small sales to try and drive subscription revenue multipliers
More baseless speculation
> More baseless speculation
Your elided quote removes the five words where I declared my views as speculation openly and in plain language. The complete sentence that you misquoted opens with that:
> I would hazard a guess
I’m perfectly content to be wrong at HN; it’s a forum where we all have opinions and people rarely restrict themselves to exclusively their own expert subjects, or else we’d all never learn anything! So I will be considering the arguments made here by others before engaging with this topic in the future.
Edit: https://news.ycombinator.com/item?id=48256417 from a few days ago apparently makes some of my points much more clearly, or at least with less hostile replies. I wish I’d found it sooner, but I didn’t realize this entire post was a dupe in time to go back through its comments in detail. Would have saved me commenting at all! Ah well.
Whelp, I’m an embedded engineering consultant and will no longer recommend these products to my customers. Or rather, I will ask them to avoid these products entirely.
AMD, you can make more money selling chips than software, but take away the entry level software and you eliminate the on-ramp. I’m not buying a license to prototype.
Advanced Marking Disaster original thread: https://news.ycombinator.com/item?id=48254309
Found the UserBenchmark alt!
Discussion from 4 days ago: https://news.ycombinator.com/item?id=48254309
Is anyone else annoyed by the click-/rage-bait style of writing?
Second: Can this software be run from Wine on Linux?
Large company again makes local decision without considering the effects outside that single product line.
I wonder how many Linux GPU sales their decision to penalize Linux on their FPGA line will cost them.
Very few, entreprise users (aka volume) will pay the license, hobbyists will pirate it if need be. AMD doesn't want to do support for the hobbyists for free, that's all.
No, AMD wants to collect some rent from people running Vivado in CI/CD environments.
Understandable. But would it be much easier to release it without promise to support. That everyone would just accept.
> I wonder how many Linux GPU sales their decision to penalize Linux on their FPGA line will cost them
Honestly? I bet the number is in the hundreds of units total.
Most people do not care that a software package they don’t use and possibly never heard of before today no longer has a free tier on Linux.
I mean, over the years, I have purchased (or advised other people to purchase) multiple Nvidia GPUs for compute workloads.
And the reason pretty much always came down to good integration between various open source software and proprietary CUDA drivers. And the assumption is that this support will continue for many years.
So, yeah, burning their existing FPGA users is a strong signal never to invest real money in their GPUs for compute workloads.
>I wonder how many Linux GPU sales their decision to penalize Linux on their FPGA line will cost them.
Not many I would guess.
considering nvidia has garbage gpu drivers in linux land and amd has pretty good ones i suspect you’re correct.
While Vivado/Vitis etc do amazing things, I challenge anyone to find a person who enjoys using them without TCL interfaces.
These tools do need attention, it's too bad there's not a better model than subscription bases like these.
Pretty sure, based on TCL base, that these tools were native Unix at some point, so the no-linux-free-beer vs windows-free-beer version are hilarious...
Ultimately one has, with so many vendor tools, a windows box somewhere so make it a remote compile machine.
Never understood why FPGA vendors do it. Do they desperately want to show software ARR to shareholders?
Smacks of collusion honestly. Maybe Microsoft offered them some kind of deal
What are some good alternatives (to Xilinx/AMD) for hobbyists / small outfits?
As far as I can tell, Lattice products are the best if you're interested in using F/OSS tools. Yosys / Icestorm supports the Lattice ICE40 and ECP5 devices. QuickLogic also apparently do some officially supported F/OSS stuff for their devices, but I don't know much about them. Think I might have ordered one of their dev boards a while back, and then promptly put it on a shelf and still haven't gotten around to touching it. :-(
The problem is, none of these devices can match the raw performance of the devices from Xilinx / Altera / Microchip / etc. So you're kinda limited if you are trying to do something that needs serious horsepower.
Always think about stuff like this, when asserting how much better AMD happens to be versus NVidia.
Folks feel outrage when companies start charging for things that were once free.
Okay, but what if you run a company whose business model no longer supports giving away free stuff? How can you transition? What would users consider less outrageous?
AMD isn't giving away free stuff. They are selling FPGA hardware. But further, the free stuff has a lot of restrictions around it which practically gear it only for university usage. And the reason they do that is they want to have new graduates have experience with their software so they can demand it from their future employers.
Basic Vivado is the bare minimum to develop for their hardware. A large amount of functionality is still locked away behind paid IP.
Most of the revenue comes from the IP cores.
A common business model for companies like this is to enable developers to learn their tools cheaply, so that when they develop something for their employer, they're more likely to reach for those tools/ecosystems and have the employer pay for those tools.
This just cuts out beginner/hobbyist FPGA devs from using industry standard tooling.
It’s still free on Windows. Your argument doesn’t have legs to walk on.
Even more than that, they still have to maintain the software to work on Linux, because they have a paid on Linux.
So if they have to keep maintaining it and offer the basic tier for free on Linux... just why? It doesn't make any sense to me.
Maybe they receive "too many" bug reports from Linux users?
And not because Linux is more buggy but because Linux is populated with people that tend to know how to make bug reports ?
You need to buy their hardware to use it. In fact there is no way to use most of their hardware without Vivado. So it's more like they are blocking you from using things you've already paid for
The software is useless without their chips and the chips cost a fortune. It's not "business model no longer supports giving away free stuff". It's just bean-counters cutting corners.
I'm guessing it works fine under Wine.
It'll be shit for paying Linux customers too as a large user base of "free"-version Linux users could spot and report bugs easily.
> Starting with the 2026.1 release
Don't upgrade. It's just that simple.
Do they offer some unique features in the new version or is it a habit to upgrade everything every day?
QoR for advanced and large designs can change wildly between versions (for better or worse.)
Yes, working with recent distros. At some point I spun up a vm because there was no way to make it work after an upgrade.
Read the article. The old version will not be supported soon. I assume that means you also won't be able to register it.
People are still using Vivado 2019 (and earlier) and ISE. The ability to use these versions isn't going anywhere.
When AMD bought Xilinx I was hoping they'd open up the software side like they (eventually) did with their GPU drivers. Looks like that isn't happening anytime soon.
It seems silly to put up SW barriers for people to use your fairly expensive HW, but what do I know.
But you just showed you have deep pockets and they think they can get you to open it again every year for the rest of time.
Xilinx was never positioned that it made sense for them to open it up. If/when it gets run into the ground by AMD short sightedness they might just open it to claim that was the plan all along...
Pretty sure this 'article' was written by an LLM, having scraped the HN discussion on here from 4 days ago. Nothing new there apart from a clickbait title and a ton of ads.
Link to my comment, so that I don't repeat myself: https://news.ycombinator.com/item?id=48256417
Kinda shitty but just a minor speed bump. Just run Windows on a VM in Linux, right?
I have specifically chosen AMD _many_ times in the past precisely because of their better linux support and more open toolchain.
This is an absolute foot-gun moment. And the gaslighting PR responses are just unacceptable. I'm very disappointed in them.
Nvidia supports their cards for many years - even quite old cards often have modern drivers.
AMD just does not see the world this way.
NVIDIA ended support for their 10xx series [1]. To be clear, AMD also moved support for their equivalent 5xxx series to legacy drivers [2], but "supports their cards for many years" doesn't hold value if both companies stopped their respective GPUs at basically the same time.
Also remember that one of those 2 companies has opensource drivers for Linux for their old GPUs, while the other doesn't (newer NVIDIA GPUs have an opensource driver but this isn't the case for the 10xx series). Users of legacy NVIDIA cards needs on Linux needs to use their old driver branches, with results that are less than optimal to say the least.
[1]: https://videocardz.com/newz/nvidia-officially-ends-geforce-g...
[2]: https://arstechnica.com/gadgets/2025/11/amd-says-that-its-no...
This is about their FPGA tooling. It has nothing whatsoever to do with GPUs.
So? I'm making a true observation about the companies. I am well aware this is about FPGA and that has nothing to do with my comment.
It is completely different. FPGA tooling is not the same as a driver for a consumer product.
A lot of the serious CUDA compute stuff is also not supported on all platforms (it's linux only, because why would you do such stuff on windows).
Your "true observation" doesn't contribute to the context of this particular topic thread which "has nothing to do with [your] comment", as you are "well aware". You should review the HN Guidelines: https://news.ycombinator.com/newsguidelines.html
[dupe] https://news.ycombinator.com/item?id=48254309
> Its comparatively smaller user base means less commercial pressure, making it an easy target to throw under the bus whenever companies feel like cutting costs or boosting profits.
Eventually the empire will strike back though.
I now marked AMD as a company that can not be trusted.
We need more indie companies in general, and cheap 3d printing for the masses. It'll be a long way to go to nanoscale perfection, but we'll have to go it - AMD but also Intel before, showed that NONE of those mega-greedy corporations can be trusted. They'll always try to do a switcheroo move. But as I stated here: the empire will strike back eventually. The Barbara Streisand effect is real.
I mean perhaps the silver lining is the projects I use are all stuck on 2022.1 for now. I wonder if this is because they want to gate usage by AI agents.
Earlier discussion: https://news.ycombinator.com/item?id=48254309
Also this site (itsfoss.com) is unusable and riddled with hundreds of ads and sets my machines fans to full blast.
At least use another credible source or go to the source instead as per the HN guidelines.
The rumor on the FPGA reddit is that they're going to walk it back.
Quote: 'The only source I can give at this time is "trust me bro"'
Hot Aisle on Twitter say they heard it directly from AMD, but I haven't seen any kind of public announcement yet.
https://twitter.com/HotAisle/status/2059706563665998317
I think AMD simply looked at the numbers and they get a lot of support requests for the free version, more than the Windows version
(of course that's the bean counter calculation without factoring in "karma")
And I kinda agree, the cost of supporting those tools on different platforms is not great
Honestly just run Wine
if "the ai" was delivering value you'd think it would be lowering both cost to develop these dev tools as well as the cost to support them. In that light the move is odd unless they don't know how to use AI to lower those costs or it cant.
Imagine if AMD focused on making their tools better instead of resorting to sleazy tactics.
Imagine if the whole industry made interoperable tools that worked on open data formats and competed on merit instead of customer lock-in.
Imagine the world we could have.
We shouldn't wait for them to get their act together, as it's in the best interest of a cartel to not have competitors, compatibility, and transparency.
It should be required after a certain amount of time that schematics and code be open sourced and that anti-walled-garden measures are prevalent so we get compatibility and extensibility right out of the box.
Incredible, behaving as if they want another CUDA situation.
They want their *own* CUDA situation. if you're doing FPGA stuff, the Xilinx hardware is good and Vivaldi is really the only way to use it. AFAIK the open-source fpga situation is pretty far behind it (please correct me if I'm wrong, I work in a place that uses Xilinx FPGAs)
> They want their *own* CUDA situation
What better way to do that than decrease availability, I ask, both rhetorically and sarcastically! CUDA-proper did well, at least partially, because it was put in front of everyone. This is taking an exclusivity angle that doesn't make much sense for ecosystem development, IMO.
I suspect this goes to show how much influence/priority B2B carries, my point is it's a mistake they've made before. The maintenance/support costs I believe they're trying to remove have paid dividends for CUDA and countless others. Anyway, a play on the tagline before I go: "Together we advance... licensing and removal of [positive] feedback loops over product development"
I seriously think the CUDA situation was set up intentionally.
Jensen Huang (Nvidia CEO) is related to Lisa Su (AMD CEO).
A decade ago, I saw a demo of some AMD skunkworks GPU datacenter tech, that could execute CUDA natively on AMD/ATI graphics cards. Initially was half speed, but having the flexibility was crazy amazing. Created a big buzz in the big iron and educational markets.
Where'd it go? Buried. You cant even find articles about it. Its a few comments on edtech datacenters.
Now look at AMD's graphics line. Where's their ROCm LLM tooling? It's a fucking joke. Its like they're intentionally sinking it for her Nvidia uncle Huang. And Su takes the cheaper CPU market and offers better features than Intel.
I'm not trying to dismiss what you're saying as a possibility (AMD's behavior in many regards over the last 15 years or so is baffling to the point that a family conspiracy feels surprisingly plausible) but Huang isn't Su's uncle.
They are "first cousins once removed" meaning that Su is the child of one of Huang's cousins. Or put another way, one of Huang's grandparents is one of Su's great-grandparents.
Honestly, my mind could never wrap around genealogy and family relations well at all. I thought they were indirect cousin/uncle , or as you say, cousins once removed.
Also my understanding of many Asian cultures is they tend to have a much more tight-knit large family structure. And doubling that is the fact they're 2 heads of world-level hardware tech companies.
And, well, there's no such thing as coincidences. Having all of this line up, and for "some reason" AMD keeps missing when they could have owned a big chunk of the market has a certain family oligopoly smell to me.
Well, but there absolutely are such things as coincidences. It is obvious that coincidences are to be expected.
Combinatorial mathematics actually says certain types of coincidence happen more often than we seem to expect intuitively (eg. the Birthday "Paradox").
I have no particular view on AMD. But any argument that includes "I don't believe in coincidences" should probably be weakened in your estimation.
That's what you get for using unfree software.
There's no free alternatives, because AMD doesn't document the bitstream format (i.e. what you need to push to the FPGA to program it to do wha you want).
This isn't why there are no free alternatives: there are for 7 Series chips. Free alternatives have terrible QoR.
I'm fairly sure the FPGA space is big enough there are alternate products for most of the offerings
Xilinx has the best silicon. Everyone else is behind. Altera is basically dead thanks Intel. Lattice is nice for low power but performance-wise they are behind. Don't know much about Microchip, but from the little I've heard their tooling is a disaster even by the standards of FPGA tooling. Then there are Gowin (not bad, but Chinglish docs and everything), Gatemate (pretty innovative and vendor-backed nextpnr support - but only one low-mid FPGA with a promise to release chiplet assemblies of it latter). And Effinix - don't know much about them, do anyone have experience?
This software seems to never have been open source/freely licensed. That's not a bait and switch. They were giving you a commercial product, for free, and now have decided not to.
It's likely a case where maintaining separate builds for the free and commercial tiers was getting complex. Often times, this kind of software requires lots of manual reviewing and adding or removing modules, and they probably decided it's just not worth it.
I don't see how that particular line of thinking applies when: 1) They continue to have a free version for Windows 2) They continue to have a version for Linux
I just can't see that cost of having a free Linux version (on top having a paid Linux version) is big?
I literally explained the thinking that the free builds on Linux aren't worth it. If you've ever shipped production software, you'd know this. Just because there's a free build available for Windows doesn't mean it costs the company $0 to release the free build. It's a lot of extra time and QA for each variant of a release. There might be many differences between the Windows and Linux builds, such as the Linux builds require proprietary 3rd party code with royalties, and they chose not to eat that cost.
There's no bait and switch. It's just people expecting things for free, as always, when this was never an open source project.
> It's a lot of extra time and QA for each variant of a release.
In Vivado, it's the same release for the free and expensive builds on both Linux and Windows. It's just a question of the installed license file/license limits.
> such as the Linux builds require proprietary 3rd party code with royalties, and they chose not to eat that cost.
This seems unlikely for a multitude of reasons.
This ignores that their revenue comes from selling chips. The software enables that. No software, no chip sales. This can only have a negative impact on their sales.
Think academic and small companies who don't pay for support opening corner case issues all the time publicly. They want none of the complex support unless you pay (reasonable imo).
And for those who forget RHEL for instance has to pay salaries to back port fixes and such and the same logic applies here.
Xilinx is/was an FPGA company until AMD bought them. Their primary revenue stream is selling chips. This is the equivalent of going back to the days of paid C/C++ compilers (anybody else remember that?).
AMD is not a good company. They stopped innovating after Intel was put down. Except, now Intel has govt backing while AMD will face significantly more competition from not only x86 but arm. Stock price says otherwise but I think they had more than enough time to catch up to Nvidia and simply refused to compete.