No, AMD Will Not Be Opening Up Its Firmware/Microcode

[starshipeleven]

I remember one of the developers (airlied?) talking about R600 microcode and how it is basically well understood what it does. But the toolchain etc. are missing to write a free replacement, and the developers are not interested in doing so.

I remember someone that said that if they publish a toolchain then it's easy to disassemble the microcode, so they don't. (I think it was bridgeman)

I'll tell you what, designing a new graphics chip takes years. So even if your claim was true (which I doubt), the competition would still be three years (give or take) behind AMD.

Huh? The competition isn't going to make a direct copycat of the AMD design as that's nonsense (they have their own already), but integrating the good parts in their own.

That can be done without an OS doesn't mean that it is not one of the tasks of an operating system.

Quit this bs already. You claimed that because some controllers in the card can write to a disk then the firmware of the card is an operating system.

I and pal666 told you that any microcontroller without an OS can do the same, which is a CLEAR indication that the ability to write to a drive is NOT a proof of the firmware becoming an operating system.

We have had drivers, DMA, and PCIe SSDs for years. So basically nothing new on the card, just a new driver?

Yes, and the drive is bolted to the card because it's used as cache for it, so it's faster if it is physically bolted to the card instead than sitting somehwere else.

As I said, I have a hard time imagining that.

This is your own problem.

And therefore it should be forbidden to say so?

No, just that you should be punished for your lack of understanding of human nature. Please read properly.

[starshipeleven]

Have you consider some code for WineHQ in order to run the Windowz Crimson driver with a VGA pass-through like those on VMs?

How in the hell can you do hardware-assisted GPU passthrough (a virtualization feature) with Wine that is not even an emulator, let alone a VM (i.e. it does not operate the hardware features necessary for this as it just translates dx to openGL plus a bunch of Windows system calls into Linux ones)?

[karolherbst]

And here is your mistake. It's not about algorithms, it's about problems. If a problem is not parallelizable, you won't found a parallel algorithm for this.

Of course a problem isn't parallelizable, because a problem isn't something you can implement at all. You can impement solutions to problems and those can be parallelizable, don't have to.

I think you missunderstand what a "problem" is at all. For every problem exist a way to solve this which is parallelizable. The question is just how efficient the solution is.

Then again, you should tell me a problem for which does not exist a parallelizable solution but before doing that, you are just blabbering. (spoiler: "hash a string with md5" isn't a problem, but a solution to one, which isn't efficiently parallelizable)

[PuckPoltergeist]

I think you missunderstand what a "problem" is at all. For every problem exist a way to solve this which is parallelizable. The question is just how efficient the solution is.

\

It seems you don't understand completely. Please inform you about P-Completeness. After this, you will understand, that there exist problems that can't be solved parallelized.

[duby229]

Have you consider some code for WineHQ in order to run the Windowz Crimson driver with a VGA pass-through like those on VMs?

What you are really asking about is something like Gallium Nine. It's good stuff. I hope to see newer DX support by similar methods sooner rather than later though.

[starshipeleven]

\

It seems you don't understand completely. Please inform you about P-Completeness. After this, you will understand, that there exist problems that can't be solved parallelized.

A quick question, does making a dedicated accelerator for that specific problem yeld better results than just trying to solve it in software?
I did read articles about how FPGAs were blasting away x86 CPUs on solving specific problems/algorithms (FPGA because making an ASIC for that would not be economically justified, of course an ASIC would have done much better), like say proteing folding (3D folding of a chain of different monomers, keeping in mind electrostatic interactions between the different monomers that compose it), and also were much easier to program than making a proper GPU compute program.

[PuckPoltergeist]

A quick question, does making a dedicated accelerator for that specific problem yeld better results than just trying to solve it in software?

Most times some specialized hardware is faster than general purpose hardware. But that's not the point. Even the most specialized hardware has to solve those problem serialized. It's about the problem itself. Not a specific algorithm for it nor specific hard- or software at all. It's the field of theoretical computer science and mathematics.

[starshipeleven]

Most times some specialized hardware is faster than general purpose hardware. But that's not the point. Even the most specialized hardware has to solve those problem serialized. It's about the problem itself. Not a specific algorithm for it nor specific hard- or software at all. It's the field of theoretical computer science and mathematics.

Yes, I was just making sure I understood this correctly.
If the problem is truly serial you can get better serialized problem solving but not turn this into parallel.

[artivision]

How in the hell can you do hardware-assisted GPU passthrough (a virtualization feature) with Wine that is not even an emulator, let alone a VM (i.e. it does not operate the hardware features necessary for this as it just translates dx to openGL plus a bunch of Windows system calls into Linux ones)?

With some effort WineHQ can run Windowz drivers like USB drivers. Also a full GPU pass-through (ready technology) can be implemented to any program, not just Virtual Box and similar programs. I don't understand why you thoughts are so limited.

[pq1930562]

bridgman :

Here's what AMD should really do:

1. Bring back the "ATI" brand. "ATI" sounds much cooler than "Radeon Technologies Group" or "AMD".
2. Stop giving your driver names like "Crimson Edition". "Crimson Edition" sounds horrible and is not cool at all.
3. Bring back cool advertisements like the following one (It was so cool, why did you stop doing such cool things?):

https://youtu.be/K0SIcFMLGeo

4. Bring back the proper Ruby, not this strange Ruby from the "Project Phoenix" CryENGINE 3 tech demo. This is the proper one:

https://youtu.be/HiszZlq3ST4

5. Make GPUs with red PCBs again
6. Fully commit to Linux / open source. Make Linux the number one priority.
7. Release everything as open source (drivers, firmware, everything, no compromises).
8. Maybe team up with Valve regarding Steam OS / Steam for Linux.
9. Stop supporting PlayStation and Xbox. Consoles and console exclusivity just ruin the PC gaming experience.
10. Rise.

Regards

You forgot less cloudy weather and more money for me.

No. But I forgot something else:

11. Remove DirectX 12 support from your drivers. Focus entirely on Vulkan instead. The world does not need yet another proprietary graphics API which only benefits Microsoft and forces consumers into buying/using Windows 10. Supporting DirectX 12 is only contributing to this. It is evil. AMD, please do good things instead.

#1 through #5 I like

Okay, good. So, are you going to arrange that to happen please? Will you do it?

May I ask what position you inherit at AMD? Are you able to pull this off?

although for #1 I would prefer to stay with AMD brand but change the logo to use ATI red. That would make our booths a lot easier to find at trade shows as well.

Hmm, no. There's an issues with this "AMD" thing. First of all, "ATI" was a well recognized brand. It was entirely okay, there was absolutely no need to abandon it. I never understood why it has been abandoned.

"ATI" was "Team Red", the direct competitor to "Team Green". It was cool. "AMD" is "Just another CPU company that also does GPUs.". Not so cool.

Also, look at the following video at the 01:00 min mark:

https://youtu.be/1DPQW0e9ufM

Even though that ape video was pretty stupid, the moment where the ATI logo appears at 01:00 min in the video is magnificent. Absolutely magnificent.

Look at that nice animation.

And listen to the sound. Wow, that sound. It makes goosebumps.

It sounds extremely similar to that sound when the Predator switches to it's visor, see:

https://youtu.be/OW1gGDbO_1U

Just wow.

That was so cool.

Is the person responsible for this still at AMD?

Please bring the ATI logo back. Please bring that sound back. Please bring that animation back.

I've only seen that animation and sound in very few videos. But it was so cool. It should be used everywhere.

Maybe even play that animation and sound everytime a game is being launched on a Radeon graphics card.

Similar to how it was done with the 3dfx animation when launching Glide API games.

By the way, the "AMD - The future is fusion" logo and animation and sound and voice at the end of the following video is also pretty cool:

https://youtu.be/K0SIcFMLGeo

Aye Amm Dee!!!

Please bring back such cool things.

It was absolutely badass!

It's good that you put #6 and #7 together, because if we did #7 then we would be unable to sell into pretty much any market other than Linux, since robust HW-supported DRM is a hard requirement for all the other markets and some of the embedded Linux market as well. Given the loss of ~95% of our market opportunities, we probably would make Linux top priority (#6) for the brief period between doing #7 and closing our doors for good.

No consumer likes DRM. Maybe let that sink in for a moment.

AFAIK we have been doing #8 for a while already.

Ah, okay, didn't realize that...

So, could you please ask Gabe to release Half-Life 3 exclusively on Linux and exclusively for Vulkan and help him achieve that using super cool ATI technology?

#9 is only practical when we have more revenue from other markets. Otherwise sales don't cover R&D costs and we go away.

If there would be no consoles, then more people would game on PC, i.e. the revenue would probably shift from the console market to the PC market. To be fair though, if you would no longer support PlayStation and Xbox, someone else will do anyway, so might be a moot point (even though it's a nice imagination).

We are working hard on #10.

Point 10 implies doing all the other points. And please also look at the freshly added point 11 above (removing DX12 support).

Regards