R600g AMD Evergreen Geometry Shaders Proposed

[TemplarGR]

You can't really expect anything from the volunteers. For example, I was asked about r300g, but since I have no such cards, what could I do?

Surely you don't expect volunteers to go purchase card X to please you.

They are not volunteers. Most are professionals, or aspire to be.

And most are paid by AMD.

[TemplarGR]

focusing on older hardware takes a lot of time for not as much return, its just economics 101.

The effort to fix DPM on r6xx is a lot harder than the effort to fix DPM support on r700 and so on down through time, as the only people who understand how dpm worked on those chips have long since moved onto something else and there are no good documents on it. Yes it might be possible to figure it out from old fglrx source code but it probably won't be that simple.

The same goes for these things, its a law of diminishing returns, I'd rather spend time fixing something for 1000s than 100s than 10s etc. and r6xx is on the long tail of things I care about anymore.

Dave.

Maybe for YOU, r6xx does not matter.

But from what i recall, DPM, UVD and other code comes from AMD, not from you or any other volunteer.

As i said, my comment was about the general attitude towards earlier generations of hardware, not specifically about you choosing to support Evergreen geometry shaders first.

[TemplarGR]

They will ultimately get customers from not caring about their old hardware. Planned obsolescence.

They won't. Customers will be avoiding their products because they hate planned obsolescence. If i wanted a more powerfull gpu or opengl4.x, i would get one. It is my right to want to stay with my hardware if it satisfies my needs.

If AMD wants to play games with driver support to force me to upgrade, i will pick Nvidia or Intel. They don't do that...

[Tobai]

I find it baffling how someone can still defend AMD's absolutely contemptible policy concerning driver support of their hardware in 2014. This isn't a criticism of the volunteers' work -- the genuine ones, those not on AMD's payroll. Fully working drivers isn't a bonus in a graphics card, it's an integral part of the product. A company that dismisses 2 or 3 year old hardware as being legacy -- even when they don't say it in so many words -- doesn't deserve anyone's money. If saying this is trolling we might as well call it a day and forget about debating anything.

[log0]

no I had no idea AMD were working on radeonsi when I started this, Red Hat didn't have a need for it either, it was just a personal side-project that I did while avoiding doing my real job.

I also didn't start the r600g code, Vadim wrote most of the initial code in his own time, I'd never have started it otherwise.

Dave.

I know a HD4850 that will be very happy to hear this.

Thanks a lot Dave and Vadim. You guys rock.

[halfmanhalfamazing]

I must be the only one in this thread who remembers

I remember. It was only 5 or so years that open source drivers were non-existent.

I remember. What it was like, the nvidia driver and fglrx was it. That's all there was.

I remember. Sticking with AGP based boards in upgrade cycles because by some miracle, ATI had released partial GPU specs for r100 and r200 class hardware. r200 was the only game in town as far as I was concerned - it had an open source driver and LOTS of support. My Radeon 8500 was like a bar of gold bullion. I had that card for a good 6-8 years for that reason, and you couldn't pry it from my cold dead hands.

Now look at what we have. A fully(with some exceptions) open graphics stack going right into brand newly released hardware.(R9) And we have a serious customer who wants their older hardware supported in the open source graphics stack.

My, how things have changed.

[halfmanhalfamazing]

The effort to fix DPM on r6xx is a lot harder than the effort to fix DPM support on r700 and so on down through time, as the only people who understand how dpm worked on those chips have long since moved onto something else and there are no good documents on it. Yes it might be possible to figure it out from old fglrx source code but it probably won't be that simple.

Bears repeating.

[_SXX_]

Maybe for YOU, r6xx does not matter.

But from what i recall, DPM, UVD and other code comes from AMD, not from you or any other volunteer.

It's okay that you unhappy with AMD drivers and customer service and I understand you. You can blame AMD and post feedback on their products quality anywhere you like, but what a point to post such messages on this forum? Do you understand that AMD employees who actually participate on this forum are NOT decision-makers?

I own HD6950 and stick to open source driver for about two years already and I'd prefer to get OpenGL 3.3 sooner and not later. I'm don't have any intention to use bugged crap called Catalyst even if it's "supported".

With such posts here you can only spoil the mood of devs who do hard and important work within limited human resources. And I totally sure that people who check news about open source AMD drivers are know situation with this hardware on Linux already.

[halfmanhalfamazing]

Actually, far more people currently use r600g than radeonSI.

Like me. I currently own two of them in two different computers.

And that's the reason why I know that all of this will get fixed. It will happen; will just take longer, as Dave said. The amount of working knowledge is reduced, so it makes sense that it might take longer. I'm not really too worried about it.

What worries me isn't Shaders and the more "common" GPU stuff, it's features like UVD that are currently held behind an iron curtain at AMD HQ. That(as it stands now) will never happen. Now that's upsetting.

[archibald]

I'm going to cast a vote for "Thank you Mr. Airlie and Mr. Girlin for your work on this.", if I ever meet you, I should like to buy you beer.