A Detailed Look At The Failed GPLGPU Open-Source GPU

[duby229]

They define what "custom design" is in the same fucking page of that pdf, slide number 5. Quit being an idiot, even your "evidence" says you're wrong.

-Full Custom ASIC- (note the CUSTOM in the name plz)
-Every transistor is designed and drawn by hand
-some stuff not relevant for us here
-Typically only used for analog portions and for very high volume parts (e.g.microprocessors) or for small parts to be used in many different designs

So to recap:

1. we have that page that says what I say too, microprocessors and other high-performance ICs are designed by hand for the most part because high volume can offset costs and high performance is required. (by hand still means making modules and assembling them with an IDE)

2."small parts to be used for many different designs" means "making a library" unless english changed very recently.

And there you go quoting me out of context and then misinterpreting the source. Argh...

So I guess, I'm going to have to find a link about the masking process so you can understand what they mean by hand drawn.

Here's a link that mentions it, but doesn't really explain why.
http://ocw.mit.edu/courses/electrica.../lecture_5.pdf

Here is a decent article that describes when very low level modification is needed and why.

http://www2.dac.com/46th/proceedings/slides/02u_2.pdf

Read it and stop this bs. Pretty much every IC is semicustom. Unless it's entirely analog.(the source you misinterpreted is really talking about place and route, part of the floorplanning and layout stage.)

[starshipeleven]

So I guess, I'm going to have to find a link about the masking process so you can understand what they mean by hand drawn.

Still more of the same.

All stuff done at the design table and not by the foundry, using advanced IDEs and libraries of pre-made things to assemble a IC design does not make the design non-FOSS (unless the library is non-FOSS, that is).

Please show me foundries that alter designs, as that's the only thing that would make printed opensource hardware non-FOSS.

And no, "hand-drawn" is still made manually at low level even in those pdfs. They state that of course in most cases you try to avoid that, but all refinements of a CPU/GPU design must be done at low level, as you cannot just "add moar cores/ComputeUnits/whatever".

[duby229]

Still more of the same.

All stuff done at the design table and not by the foundry, using advanced IDEs and libraries of pre-made things to assemble a IC design does not make the design non-FOSS (unless the library is non-FOSS, that is).

Please show me foundries that alter designs, as that's the only thing that would make printed opensource hardware non-FOSS.

And no, "hand-drawn" is still made manually at low level even in those pdfs. They state that of course in most cases you try to avoid that, but all refinements of a CPU/GPU design must be done at low level, as you cannot just "add moar cores/ComputeUnits/whatever".

That's exactly what I'm saying because it's exactly what happens. IC's are developed using proprietary tools with proprietary standard cells, interpreted and fabricated by proprietary processes. Whether you like it or not or agree with it or not pretty much every asic is semicustom.

[starshipeleven]

IC's are developed using proprietary tools

True, would not be an issue for FOSS hardware

with proprietary standard cells,

True, there are FOSS cells for FOSS hardware so again a non-issue

interpreted

There is no interpretation. "write a transistor here connected to these lines" does not require any interpretation. As the end result of the design phase is a big huge map of transistors and electrical connections between stuff. Fabs aren't compilers, fabs are printers.

and fabricated by proprietary processes.

True, is not an issue for FOSS hardware as no change to the design is made, fabs do their best to keep the product as close to the design as possible, not to optimize or do shit. Anything that deviates from design is manufacturing error and is labeled as such.

[duby229]

True, would not be an issue for FOSS hardware
True, there are FOSS cells for FOSS hardware so again a non-issue
There is no interpretation. "write a transistor here connected to these lines" does not require any interpretation. As the end result of the design phase is a big huge map of transistors and electrical connections between stuff. Fabs aren't compilers, fabs are printers.
True, is not an issue for FOSS hardware as no change to the design is made, fabs do their best to keep the product as close to the design as possible, not to optimize or do shit. Anything that deviates from design is manufacturing error and is labeled as such.

The part that I bolded is simply wrong. Read the links I posted again.

Also I think you have a fundamental misunderstanding of what free and open source actually is. They definitely aren't the same thing. Something free as in freedom must be completely open source, but not all open source is freedom. That's what I'm saying you can call it open source, but you can't call it free.

EDIT: Seriously, RISC-V itself has giant arm core macros in it. Really.

[starshipeleven]

The part that I bolded is simply wrong. Read the links I posted again.

Those links deal with development, creating the map with all the transistors in their place. You can do all sorts of things depending on what you want and how much you can pay the developing team, but in the end you get a full map (and there are also some pictures of some of these) you can send to print to the fabs that will print your stuff.

Also I think you have a fundamental misunderstanding of what free and open source actually is.

https://www.gnu.org/philosophy/free-sw.html
-------------
“Free software” means software that respects users' freedom and community. Roughly, it means that the users have the freedom to run, copy, distribute, study, change and improve the software. Thus, “free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech,” not as in “free beer”. We sometimes call it “libre software” to show we do not mean it is gratis.
------------
If we change "software" with "hardware design", that is what FOSS hardware is.

Is the design available to be "run" (printed in an ASIC or used in a FPGA)? Yes.
copy/distribute? yes.
study? yes.
change and improve? yes. This point does not mean that FOSS software should drop a full toolchain and IDE and university degree on top of everyone, just that the source itself is available for processing. The same happens with open hardware designs.

and again:

----------------------
A program is free software if the program's users have the four essential freedoms:

• The freedom to run the program as you wish, for any purpose (freedom 0).
• The freedom to study how the program works, and change it so it does your computing as you wish (freedom 1). Access to the source code is a precondition for this.
• The freedom to redistribute copies so you can help your neighbor (freedom 2).
• The freedom to distribute copies of your modified versions to others (freedom 3). By doing this you can give the whole community a chance to benefit from your changes. Access to the source code is a precondition for this.

----------------------------------

This too is allowed by a open hardware design.

[duby229]

Those links deal with development, creating the map with all the transistors in their place. You can do all sorts of things depending on what you want and how much you can pay the developing team, but in the end you get a full map (and there are also some pictures of some of these) you can send to print to the fabs that will print your stuff.

https://www.gnu.org/philosophy/free-sw.html

-------------
“Free software” means software that respects users' freedom and community. Roughly, it means that the users have the freedom to run, copy, distribute, study, change and improve the software. Thus, “free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech,” not as in “free beer”. We sometimes call it “libre software” to show we do not mean it is gratis.
------------
If we change "software" with "hardware design", that is what FOSS hardware is.

Is the design available to be "run" (printed in an ASIC or used in a FPGA)? Yes.
copy/distribute? yes.
study? yes.
change and improve? yes. This point does not mean that FOSS software should drop a full toolchain and IDE and university degree on top of everyone, just that the source itself is available for processing. The same happens with open hardware designs.

and again:

----------------------
A program is free software if the program's users have the four essential freedoms:

• The freedom to run the program as you wish, for any purpose (freedom 0).
• The freedom to study how the program works, and change it so it does your computing as you wish (freedom 1). Access to the source code is a precondition for this.
• The freedom to redistribute copies so you can help your neighbor (freedom 2).
• The freedom to distribute copies of your modified versions to others (freedom 3). By doing this you can give the whole community a chance to benefit from your changes. Access to the source code is a precondition for this.

----------------------------------

This too is allowed by a open hardware design.

The two parts I bolded are one and the same and is specific to the foundry you choose. So the answer is obviously no. It does not fit the definition for libre or freedom. And again you can design an open source project that uses proprietary standard cells but that doesn't fit the definition either.

[starshipeleven]

The two parts I bolded are one and the same and is specific to the foundry you choose.

No it is not. You did not show anything that supports that.

And even if it were, it would be like saying that a program written in python 2.x that cannot work with python 3.x without some tweaking (python 3.x is not totally retrocompatible) is not FOSS just because you aren't free to choose whatever compiler you want.

There are also cadres of functions and features that get dropped (for good reasons) from compilers, does that make all projects that need them non-FOSS because you don't have the freedom to use them with the newest compilers? No. They are still FOSS.

And again you can design an open source project that uses proprietary standard cells but that doesn't fit the definition either.

No you (usually) cannot. There are (usually) NDAs that prevent it. This is a non-issue as there are foss cells anyway, like for RISC-V, but also in other open-hardware projects.

[FreedomHash]

I had to join to comment on this ridiculous conversation. Couldn't help myself.

I don't know much about the IC hardware design process. But probably not less than either of you. This seems one of those "debate by real-time learning" things, where you are both frantically googling.

That said, duby229 clearly and unambiguously wins this "debate", hands down.

Mainly, because starshipeleven has gone out of his way to:

1. Be rude, insulting, bullying, and generally unprofessional if not outright unhinged.
2. Misrepresent duby229's argument.
3. Make circular arguments without really adding anything.
4. Generally act like a know-it-all when it's obvious he doesn't know it all.
5. Make the internet a slightly worse place.

Whatever legitimate arguments starshipeleven might have had, are washed away by that. I don't know who is actually, factually right, and I'm not basing any personal decisions on this proclamation of the obvious winner, but there is an obvious winner. And if I had to wager, I'd wager that duby229 is actually, factually right.

I'd really like to hear what qualifications both of you have for making your claims. If your answer seems suspiciously relevant and expert, I'll ask the same question for one or more other unrelated posts where you also seem to be trying to exude knowledge and expertise. (Or maybe not. I am lazy and have no dog in this fight. OTOH no one likes a know-it-all, least of all me!) And if you really do have the expertise you might claim, you won't bat an eye at that.

[starshipeleven]

I don't know much about the IC hardware design process. But probably not less than either of you. This seems one of those "debate by real-time learning" things, where you are both frantically googling.

If you look his post history you can see this is just a part of a larger flamewar between me and him, where he claims bs and I (and others) bash him over either his bs or the total lack of proofs for his claims.

That said, duby229 clearly and unambiguously wins this "debate", hands down.

Rules of debate apply on debates, this one wasn't.
My main goal was stfu-ing him because he was posting bs, not trying to convince him with a fair discussion as this is not possible, he just repeats his bs over and over and over thinking it makes that right, regardless of any evidence you post.
I just used his own standard tactics against him.

And if I had to wager, I'd wager that duby229 is actually, factually right.

Sorry, but I'll have to call you moron if you do so. Slightly better manners or abidance to rules of conduct don't make one's arguments right.
Evidence is the only thing that makes you right.
Sure in a debate game like those you do at school he would have won the game, but this is not a game, nor a debate.

If you encounter a similar situation, the most logical course of action is looking at proof. Neither of us has convincing proof? Cool, neither is right. Case closed.

I'd really like to hear what qualifications both of you have for making your claims.

Huh? You aggravate me with rules of debate then ask about qualifications, as if qualifications matter in a debate? You know what is an "appeal to authority" logical fallacy?
You should want proof, not authority. Also because anyone can fake authority on the internet.