The official replies are addressing questions that nobody has asked. The main issue is why Linux support is being removed from the Basic tier while Windows is still allowed.
To grow the ecosystem, AMD needs more people working on their hardware. Restricting Linux will only alienates students, hobbyists, and devs who want to adopt AMD tech.
The official replies started off by addressing ... the "unacceptable abusive behavior towards AMD". The most important thing here is obviously to ask people not to use such hurtful words as "disgraceful" towards poor little AMD...
Answering the actual question seems not a high priority
Some people, including the management of most big corporations, claim that verbal insults, which do no actual physical harm to anyone, are "unacceptable abusive behavior", while the actions that do physical harm to others, e.g. by tricking or forcing them to pay an extra part of their hard-earned money for things that should not have been paid, because they had already been paid in another form, instead of using that money for worthy purposes, are not "unacceptable abusive behavior".
Obviously, I believe that a decision like that made by AMD now is a much more "unacceptable abusive behavior" than any kind of verbal insult ever known to mankind.
This kind of decision is a masked price rise of the AMD FPGAs that applies only to small businesses and individuals, while the big quasi-monopolistic companies are not affected, which will make competing with them even more difficult.
When they do not have any justifiable answer, or don't want to answer, but need to keep the facade on, they'll sidestep and tell you how hard they are working on something, and how many unrelated things they've archived.
- A regular tactic used by our former autocratic ruler, or most corrupted people
On the other hand - this is now an opportunity for Linux community to show that they are actually able to fund development of software for their platform, right?
Many HNers promised to pay if developers bring their software to Linux - will that actually happen?
The only reason why the "Linux community" cannot create adequate FPGA design tools is that the vendors like AMD refuse to document the necessary details of their products.
A few old AMD FPGAs have been reversed engineered, e.g. some ARTIX-7, so for them there is no need for the rather bad AMD tools, but for most AMD formerly Xilinx FPGAs it is impossible to create better tools for lack of documentation.
As long as AMD refuses to provide the technical documentation required to use their products, it should have been a legal obligation to at least provide basic tools that allows the buyer of such products to actually use "FPGAs", i.e. to "field-program" them, as the name of the sold product claims.
Like many other FPGA developers, I could write myself better FPGA development tools than what AMD provides, if I had access to the complete FPGA technical documentation to which only a few big companies have access, a restriction whose only possible purpose is to prevent competition in the FPGA market.
If AMD had documented the exact format of the bit stream required to program each model of their FPGAs and the complete timing consequences of each synthesis choice, nobody would need any FPGA simulation or synthesis tool provided by AMD in Vivado.
Vivado already supports Linux, the development is supported by very large customers that put FPGAs in cars, [REDACTED], and other kinds of objects that crash into other objects.
We've had good experiences with Lattice parts. Their software tools are free for all of their basic chips. They only charge for licensing when you use the higher end SKUs with SerDes. Example, you can use and develop on an ECP5 or Certus using their free license, but then you need a paid license to work on ECP5-5G or CertusPro chips.
They're not perfect, but they're better to work with than Xilinx. Also, their datasheetd are better than Xilinx in my experience.
I’ve spent several hundred thousand on Xilinx FPGAs yet they nickel and dime me for licenses. It’s not the cost that’s a problem—-it’s the hassle of making a PO for a license to set up new computers, set up CI, hiring new teammates, setting up for interns/students. Xilinx has continued to go downhill since their acquisition by AMD.. it used to feel like it was run by engineers who understood their customers, now it seems to be getting taken over by the MBA crowd who only understands pinching pennies and chiseling their own loyal customers
I can get parts, they're part of a BOM that gets approved, but getting POs approved for software is a pain in the ass. Been considering switching next gen stuff to microchip.
It's really unfortunate that FPGA development is still stuck in the 90s. The incentives between IP owners and hobbyists are so misaligned that I don't see the possibility of this ever improving.
The market is full of dark patterns, and vendors like AMD/Xilinx can pull shitty moves like what OP highlighted, knowing there is no decent alternative (Altera is another disaster). Lattice had the opportunity to fully embrace opensource toolchain and try to disrupt from the bottom, but they seem stuck in the middle, not wanting to commit one way or another.
I'm grateful to SymbiFlow, and IceStorm and others, even though they obviously lack support for proprietary hardware features.
For reverse engineering, you still need access to the FPGA tools provided by the vendor, to see what changes in the bitstream when you change the design.
If the bitstream is encrypted, you will not see the changes, so the only way is to reverse engineer the Vivado executables.
You do not need only the bitstream, but you also need a huge amount of timing parameters. In theory, they could be obtained by fuzzing, but that would require a huge amount of executions of the Vivado tools. So again the most plausible method is to reverse engineer the Vivado executables, to get the timing parameter database.
In some countries that should be legal, as such reverse engineering might become the only way to use the AMD FPGAs that one buys legally.
The difficult part is the place and route algorithm, not the bitstream. The proprietary ones already take quite a long time to solve: I regularly have 12-24h runs. Perhaps an open source one could do better? But it's not quite as straightforward as reverse engineering a proprietary bitstream.
As someone actively working on nextpnr support for a fairly new FPGA architecture, it really is amazing that we have something like that in the open source world.
YosysHQ are one of my favorite companies to exist.
When I first started doing chip design my boss paid more for tools per year than he paid me ... now days open source tool chains are leaping ahead ... I don't need a boss (or VCs) in order to design chips
How is it sustainable for AMD to maintain their software on Linux for free? Would you maintain your own Linux software (and its distros) for $0?
I see no problem with monetizing Linux users. If I am monetizing Windows and macOS users, there should be no exceptions towards Linux especially as Linux support is always ill defined (there are hundreds of distros to support and test.)
1: The software is not free. There is what essentially amounts to a free trial. This free trial used to support Windows and Linux. Now the free version only supports Windows, only the paid tiers work on Linux.
2: The software is what amounts to a hardware-specific compiler/IDE. AMD sells the hardware, with healthy margins. Asking "how is it sustainable for AMD to maintain [Vivado] .. for free" is the same as asking, "how is it sustainable for AMD to maintain their OpenGL drivers for free". They have a solid revenue stream from hardware sales that's enabled by the software.
3: Maintaining a free Linux version is close to 0 additional cost. They already need to maintain a free tier because they provide that to Windows, they already need to maintain Linux support because they provide that for the paid tiers. The only extra maintenance would be whatever edge case bugs occur only on the free tier and only when compiled for Linux.
First of all, you already pay for the FPGAs, and the only reason why you pay for them is that they are "Field-Programmable" GAs. To be able to use the product that you own, as advertised, you MUST have the AMD software tools, because they refuse to provide the technical documentation that would allow the use of FPGAs without vendor-provided tools.
It is abusive to request an additional big payment in order to use the bought product as intended. This additional payment for the FPGA programming tool is negligible for big companies, which also get great discounts in the price of the FPGAs they buy, but it hurts any small companies and individuals who want to use FPGAs.
These kind of policies never increase in any way the revenue of a company like AMD but they ensure that any market where such policies are frequent become dominated by a few quasi-monopolies, instead of having a healthy competition that keeps prices low for computers, as it existed in electronics until around a quarter of century ago.
Their FPGA development software is not an independent product, but it is a part of the FPGAs they are selling, like the boxes in which such FPGAs are packaged.
Your claim that they get $0 for their software is as ridiculous as the claim that Intel can no longer sell boxed CPUs, because they get $0 for the cardboard and plastic packages of their CPUs.
For now, only the Linux version of the FPGA tools has been discontinued, the free and worse Windows version still exists, so what you say in the last version of your comment is still wrong, because the Windows users are not monetized, yet.
You'd think removing friction on the software side for someone who already bought their hardware would be in their interest. Especially for students and hobbyists, who will want use what they already know once they enter the industry.
The official replies are addressing questions that nobody has asked. The main issue is why Linux support is being removed from the Basic tier while Windows is still allowed.
To grow the ecosystem, AMD needs more people working on their hardware. Restricting Linux will only alienates students, hobbyists, and devs who want to adopt AMD tech.
- From long term AMD user
The official replies started off by addressing ... the "unacceptable abusive behavior towards AMD". The most important thing here is obviously to ask people not to use such hurtful words as "disgraceful" towards poor little AMD...
Answering the actual question seems not a high priority
Some people, including the management of most big corporations, claim that verbal insults, which do no actual physical harm to anyone, are "unacceptable abusive behavior", while the actions that do physical harm to others, e.g. by tricking or forcing them to pay an extra part of their hard-earned money for things that should not have been paid, because they had already been paid in another form, instead of using that money for worthy purposes, are not "unacceptable abusive behavior".
Obviously, I believe that a decision like that made by AMD now is a much more "unacceptable abusive behavior" than any kind of verbal insult ever known to mankind.
This kind of decision is a masked price rise of the AMD FPGAs that applies only to small businesses and individuals, while the big quasi-monopolistic companies are not affected, which will make competing with them even more difficult.
When they do not have any justifiable answer, or don't want to answer, but need to keep the facade on, they'll sidestep and tell you how hard they are working on something, and how many unrelated things they've archived.
- A regular tactic used by our former autocratic ruler, or most corrupted people
On the other hand - this is now an opportunity for Linux community to show that they are actually able to fund development of software for their platform, right?
Many HNers promised to pay if developers bring their software to Linux - will that actually happen?
What you say is ridiculous.
The only reason why the "Linux community" cannot create adequate FPGA design tools is that the vendors like AMD refuse to document the necessary details of their products.
A few old AMD FPGAs have been reversed engineered, e.g. some ARTIX-7, so for them there is no need for the rather bad AMD tools, but for most AMD formerly Xilinx FPGAs it is impossible to create better tools for lack of documentation.
As long as AMD refuses to provide the technical documentation required to use their products, it should have been a legal obligation to at least provide basic tools that allows the buyer of such products to actually use "FPGAs", i.e. to "field-program" them, as the name of the sold product claims.
Like many other FPGA developers, I could write myself better FPGA development tools than what AMD provides, if I had access to the complete FPGA technical documentation to which only a few big companies have access, a restriction whose only possible purpose is to prevent competition in the FPGA market.
If AMD had documented the exact format of the bit stream required to program each model of their FPGAs and the complete timing consequences of each synthesis choice, nobody would need any FPGA simulation or synthesis tool provided by AMD in Vivado.
Vivado already supports Linux, the development is supported by very large customers that put FPGAs in cars, [REDACTED], and other kinds of objects that crash into other objects.
This is just hurting students and hobbyists.
We've had good experiences with Lattice parts. Their software tools are free for all of their basic chips. They only charge for licensing when you use the higher end SKUs with SerDes. Example, you can use and develop on an ECP5 or Certus using their free license, but then you need a paid license to work on ECP5-5G or CertusPro chips.
They're not perfect, but they're better to work with than Xilinx. Also, their datasheetd are better than Xilinx in my experience.
Give Lattice a look for your next project.
I’ve spent several hundred thousand on Xilinx FPGAs yet they nickel and dime me for licenses. It’s not the cost that’s a problem—-it’s the hassle of making a PO for a license to set up new computers, set up CI, hiring new teammates, setting up for interns/students. Xilinx has continued to go downhill since their acquisition by AMD.. it used to feel like it was run by engineers who understood their customers, now it seems to be getting taken over by the MBA crowd who only understands pinching pennies and chiseling their own loyal customers
Tbh, I think they should just charge for the chips and keep the software free.
This x10000
I can get parts, they're part of a BOM that gets approved, but getting POs approved for software is a pain in the ass. Been considering switching next gen stuff to microchip.
Microchip also wants money for license to their design suite.
The parts were considering are available under their free tier IIRC
Link to AMDs description of the new pricing being criticized: https://www.amd.com/en/products/software/adaptive-socs-and-f...
I love the way they buried the “we are no longer supporting an entire operating system” in a small missing tick, half way down the page…
Oh, no, it's sillier than that.
They do still support Linux... but only if you give them money.
This might be a good opportunity for Microsoft to bring back their 'total cost of ownership' advertising campaign.
You don't 2.5x your stock price in two months by underextracting value.
I’m working in education and will change to other vendors in the near future. That means all my students will do so as well.
Windows cannot provide feature parity for workloads that require cross compiling, AMD could at least support RHEL like the old days.
Good news for FOSS FPGA toolchains, I suppose. Eg https://f4pga.org/ for some kind of umbrella project.
It's really unfortunate that FPGA development is still stuck in the 90s. The incentives between IP owners and hobbyists are so misaligned that I don't see the possibility of this ever improving.
The market is full of dark patterns, and vendors like AMD/Xilinx can pull shitty moves like what OP highlighted, knowing there is no decent alternative (Altera is another disaster). Lattice had the opportunity to fully embrace opensource toolchain and try to disrupt from the bottom, but they seem stuck in the middle, not wanting to commit one way or another.
I'm grateful to SymbiFlow, and IceStorm and others, even though they obviously lack support for proprietary hardware features.
This is terrible news for university users trying to professionalize their FPGA development with CI/CD. Which is probably the point of the change.
I wonder how good LLM agents are at reverse engineering FPGA bitstreams...
I want a robust open-source ecosystem where anyone can take my hardware projects and modify them without needing to deal with licensing friction.
For reverse engineering, you still need access to the FPGA tools provided by the vendor, to see what changes in the bitstream when you change the design.
If the bitstream is encrypted, you will not see the changes, so the only way is to reverse engineer the Vivado executables.
You do not need only the bitstream, but you also need a huge amount of timing parameters. In theory, they could be obtained by fuzzing, but that would require a huge amount of executions of the Vivado tools. So again the most plausible method is to reverse engineer the Vivado executables, to get the timing parameter database.
In some countries that should be legal, as such reverse engineering might become the only way to use the AMD FPGAs that one buys legally.
The difficult part is the place and route algorithm, not the bitstream. The proprietary ones already take quite a long time to solve: I regularly have 12-24h runs. Perhaps an open source one could do better? But it's not quite as straightforward as reverse engineering a proprietary bitstream.
That's why nextpnr exists :)
https://github.com/YosysHQ/nextpnr
As someone actively working on nextpnr support for a fairly new FPGA architecture, it really is amazing that we have something like that in the open source world.
YosysHQ are one of my favorite companies to exist.
When I first started doing chip design my boss paid more for tools per year than he paid me ... now days open source tool chains are leaping ahead ... I don't need a boss (or VCs) in order to design chips
Somewhere in reverse-engineering-land is the desire to figure out undocumented hardware blocks. I’m not disagreeing about PNR here.
Well Gowin here I come I guess
I took Efinix. They have cute FPGA with memory in one package. It saves me lots of time for routing the board.
What’s next? Take away mouse support in the free tier? You could these fucking cretins with GPT2 and the company would flourish.
How is it sustainable for AMD to maintain their software on Linux for free? Would you maintain your own Linux software (and its distros) for $0?
I see no problem with monetizing Linux users. If I am monetizing Windows and macOS users, there should be no exceptions towards Linux especially as Linux support is always ill defined (there are hundreds of distros to support and test.)
So much wrong with your comment.
1: The software is not free. There is what essentially amounts to a free trial. This free trial used to support Windows and Linux. Now the free version only supports Windows, only the paid tiers work on Linux.
2: The software is what amounts to a hardware-specific compiler/IDE. AMD sells the hardware, with healthy margins. Asking "how is it sustainable for AMD to maintain [Vivado] .. for free" is the same as asking, "how is it sustainable for AMD to maintain their OpenGL drivers for free". They have a solid revenue stream from hardware sales that's enabled by the software.
3: Maintaining a free Linux version is close to 0 additional cost. They already need to maintain a free tier because they provide that to Windows, they already need to maintain Linux support because they provide that for the paid tiers. The only extra maintenance would be whatever edge case bugs occur only on the free tier and only when compiled for Linux.
First of all, you already pay for the FPGAs, and the only reason why you pay for them is that they are "Field-Programmable" GAs. To be able to use the product that you own, as advertised, you MUST have the AMD software tools, because they refuse to provide the technical documentation that would allow the use of FPGAs without vendor-provided tools.
It is abusive to request an additional big payment in order to use the bought product as intended. This additional payment for the FPGA programming tool is negligible for big companies, which also get great discounts in the price of the FPGAs they buy, but it hurts any small companies and individuals who want to use FPGAs.
These kind of policies never increase in any way the revenue of a company like AMD but they ensure that any market where such policies are frequent become dominated by a few quasi-monopolies, instead of having a healthy competition that keeps prices low for computers, as it existed in electronics until around a quarter of century ago.
Their FPGA development software is not an independent product, but it is a part of the FPGAs they are selling, like the boxes in which such FPGAs are packaged.
Your claim that they get $0 for their software is as ridiculous as the claim that Intel can no longer sell boxed CPUs, because they get $0 for the cardboard and plastic packages of their CPUs.
For now, only the Linux version of the FPGA tools has been discontinued, the free and worse Windows version still exists, so what you say in the last version of your comment is still wrong, because the Windows users are not monetized, yet.
They are selling hardware.
You'd think removing friction on the software side for someone who already bought their hardware would be in their interest. Especially for students and hobbyists, who will want use what they already know once they enter the industry.
Maintaining support for Windows is free?
"free" as in what? Time?
There is always someone paying. Linux should be no different.