Valve Engineer Fixes Massive Performance Issue For RADV Driver With AMD FSR2

[finalzone]

Fuck you AMD for refusing to contribute directly to RADV and duplicating work!
After you already refuse to give us a control panel like for Windows (of course with all the drivers improvements needed for that) because you don't have the time / resources for that!
But FFS you have time for all this duplicated work!!!

Congratulations and many thanks to Valve!
No wonder why I still buy games from you even if I don't like DRM and centralized stores that can take my access away any time they want from what I have bought!

AMD already recommended RADV for consumer markets running Linux operating systemsand this community have AMD contributors (some who posted in this forum) working on MESA. Any more complain?

[ahrs]

of course, if your continued employment hinges on the open driver team not reaching your performance, then you don't find it worthwhile to mention...

Do you have any proof of any collusion happening internally at AMD (if you do, maybe you should leak it to somebody like the Lunduke Journal…) or are you just speculating?

[ezst036]

Need spellcheck so that there isn't such a gpaing hole in the written article.

[NeoMorpheus]

It wouldnt be a proper AMD article without the haters trashing AMD.

[mphuZ]

Fuck you AMD for refusing to contribute directly to RADV and duplicating work!

AMD doesn't owe you anything. And it's not AMD that duplicates the work, but the RADV developers.

After you already refuse to give us a control panel like for Windows (of course with all the drivers improvements needed for that) because you don't have the time / resources for that!

In order for Adrenalin to make sense, Linux and the amdgpu driver must be able to perform the functions (under the hood) that Adrenalin provides. Otherwise, it turns out that there is a control panel, but the banal color management (or any other function) works with errors.

[Rauros]

AMD doesn't owe you anything. And it's not AMD that duplicates the work, but the RADV developers.


In order for Adrenalin to make sense, Linux and the amdgpu driver must be able to perform the functions (under the hood) that Adrenalin provides. Otherwise, it turns out that there is a control panel, but the banal color management (or any other function) works with errors.

What is that AMD doesn't owe you anything mean? lol.
No, they owe me good service because I literally paid for that. They are not doing a charity work. I won't go bother open source developers about the features I want them to add since they literally don't owe me anything. But AMD does owe me good service. So as a paid customer I ask them to not duplicate work and start directly contribute to RADV.

[stormcrow]

I don't have a single game that uses FSR 2 (in Linux), just v1. But I am seeing rather dramatic improvement in performance in Fedora 41 across the board versus MXLinux & Fedora 40. I've yet to try out some of my more demanding games, but I'm already impressed with the more up-to-date stack in general. Remains to be seen how many papercuts I encounter however. Especially when it comes to other X.org dependent programs which is most of the programs and games I use.

Ryzen 5 5600X
AMD RX 6650 XT
F2FS for root & home on an Intel NVMe SSD (yes, a real one from before Intel sold the product line off)

Haven't tried Blender/HIP yet, but I'm not hopeful since it didn't work on Fedora 40 and there's nothing that's changed to give me any reason to believe it works in 41 on unofficial GPUs.

Edit to add: Please note, don't use F2FS on a system without a UPS or a laptop with a battery that can hold a charge. It's designed for systems that never shutdown in an unsafe way.

[smitty3268]

it's amazing AMD has a driver team that is writing their proprietary driver, that does not communicate with the team that writes the open driver, which has to reverse engineer and find out all the things themselves.

Seems like it can only stem from something like paranoid separation of proprietary information between closed and open drivers, extending so far that not even basic hints at what is the problem don't transfer over.

One team works for AMD, and the other works for Valve. I'm not sure why it would be surprising that there's not much communication between the two, they are completely different companies.

[mphuZ]

What is that AMD doesn't owe you anything mean? lol.
No, they owe me good service because I literally paid for that.

You buy their hardware. And you should use the software that they offer. Everything else is your dreams and wet fantasies.

If you paid for their hardware, then this does not mean that they should gallop and hire engineers who will simultaneously develop (as you say, duplicate) another Vulkan driver.

I ask them to not duplicate work and start directly contribute to RADV.

You definitely lost touch with reality. No one will waste time and pay a lot of money to engineers to maintain a piece of driver code. If you like it, support it. When it is profitable or when the rest of the actors (Valve, Google, etc.) come to an agreement with AMD, then there will be official support for RADV (or vice versa - AMDVLK)

[Mitch]

This change has no effect on the FSR2 upscale pass itself, that part already worked fast without problems before. It's just happened to be that the FSR2 demo scene exposed this particular optimization opportunity, but it ran slow with or without even enabling the FSR upscale itself.

Do you think the optimization would be likely to be more noticeable outside that demo? Such as in videogames? Just curious