Radeon™ Software Adrenalin 22.2.2 Download & Discussion

Nice conspiracy theory, but no.

If AMD would just implement any form of multithreading and command lists in their driver that does neither, all d3d11 titles would benefit, it has nothing to do with the Hardware capabilities at all.

DXVK emulates both and this provides improvement to titles running on AMD's inferior d3d11 driver by skipping it for vulkan

 

Sorry, but AMD (IMO) is not obligated in implementing any form of command lists for an API that was replaced with DX12. Regardless, I've already provided proof of AMD's competitor involvement in DX11 making your suggestion moot.

But as you said why should they when DX11 can be skipped for vulkan. Which is providing further proof that the game should have used Vulkan from the get go.

LOL, this is comical.

 

This stinks of an apologist settling for poor hardware support when its not necessary.

D3D11 games are not going away any time soon, not everything and not every studio will find the cpu benefits of d3d12 compelling enough to relearn and retool.

Did you miss the part where i said DXVK emulates the capability?

Did you also know that command lists for DX11 are almost exactly feature equivalent in D3D12?

The Kinetica engine has always had a D3D renderer for on PC debugging and testing of the PS games, it was faster and cheaper to adapt what they already had for D3D11 than redoing everything in 12.

 

They released these a day too early, it's not quite 22-02-2022

 

Lets complain about that, jk

Anyway im expecting new drivers in a week before shadow warrior 3 and elex 2 maybe with a fix for polaris and lighning issue in cyberpunk.

 

That's not how the versioning works.
It's Year-Month-Revision.

22 2022
2nd month
revision 2

 

I did some tests to check if the SAM changes made in 22.2.2 affected on performance on my hardware. It has no effect on my computer but i'm sharing detailed results:
https://docs.google.com/spreadsheets/d/1_IYLApMoXUDH-y9sJrquFxiduhWfdIf2CZMQtCIEyPQ/edit?usp=sharing

 

What's comical is you, both in this thread and in your HRTF audio thread, and your contempt for other members of this forum. @Astyanax happens to be correct in this.

DXVK Async converts DX11 calls into Vulkan API calls, and adds Async support to better match AMD's parallel pipeline setup. (DX11 is optimized for NVIDIA's single pipeline architecture) This allows for improved multithreading support, usually resulting in improved lower end FPS, lowered overhead, and in some games better max fps. Combined with SAM and this can go up further with faster graphical data transfers since you can move more memory data per cycle.

You only make yourself look like the clown when you spew bad information and act rude to those who disagree with you. You did this a lot in the last 22.2.1 driver thread as well. This forum isn't a contest, and your elitist attitude is a very immature "gamer" attitude. Enough. I'm actually trying to help you realize how nasty you come across in your posts.

 

Jet pack actually worked on NBA Live 14 - 16, Dark Souls and Plants Vs Zombies. Just because you never heard of them doesn't really mean anything.

 

This is the wrong context of the conversation. And you later contradict yourself below. Yet are not able to provide detail on what it is he's suppose to be correct on as it relates to why he responded to my post. It's funny how you bring up HRTF. You refuted the claims yet never had tried it for yourself until another member used the software version of Wave NX to see that it was true.

And here is the contradiction. I clearly stated in my prior post that DX11 was optimized for them. And if you have to mod a vulkan api they could have simply used Vulkan instead of DX11. So you are not making any sense there. Because the truth of the matter is simple. If you have to use that just to play the game on Radeon then DX11 hampers Radeon users. The developer should have been neutral and used either DX12 or Vulkan instead of a DX11 which was replaced so long ago. Because as I also pointed out in my prior post when Vulkan was used didn't require a revised AMD driver in order to see the performance improvements. <--That is the topic of conversation. Not you trying to cheerlead and gaslight a disgruntle post that side steps that conversation.

You are the one who failed to clarify your position on what your hostile, name calling post towards me is about. To agree to what I posted earlier to only call me names afterwards makes your response trolling in nature. You really should take your own advice and realize this isn't a contest.

I have provided my input to the topic at hand without the use of name calling and gas lighting. You can either agree or disagree but when you resort to the aforementioned just to respond only shows poor decorum on your part. You are only projecting what you are trying to falsely claim I am doing. And is a form of gas lighting when I had no prior conversation with you until your nasty, naming calling post towards me. Which deflects from the conversation at hand regarding dx11 which is obsolete now.

You should take your own advice. And realize that you are not the only one here with a view point/opinion. Nor is it with those you agree with.

 

And you're wrong.

Even the worst optimized piece of crap game that uses d3d11 runs better on Nvidia without any sort of targetted optimizations because the driver does most of the work,

AMDs D3D11 driver only got into this state (after having one of th best D3D9 drivers) because of software engineers being let go during a gross misstep of management by Hector Ruiz and continued on with Dirk Meyer

 

And you are wrong.
This is all conjecture on your part. As it was already posted nvidia intro'd DX11 during gamefest in 2008. Which I already posted a page or so ago.

Which is a distraction from the fact that you proved that the developer should have used vulkan instead of dx11. Because with existing drivers there is a performance improvement for Radeon users. I even posted about it herehttps://forums.guru3d.com/threads/radeon™-software-adrenalin-22-2-2-download-discussion.441874/page-4#post-5994715.

For you to argue with me that AMD should have went back and catered new drivers for "you" for DX11 when Vulkan works better with existing driver support is simply a absurd notion. Which is why we should continue to move forward and make sure that AMD/Radeon continues innovating current Vulkan/DX12 support.

 

anything you post is regurgitated theory and gaslit propoganda to continue a flailing mentality that everyone is out to get AMD and factors around unfounded myths of sponsorship being intended to reduce performance on AMD when the reality tends to highlight developers making bad decisions to cheat around raster bottlenecks or simply making a computationally intense game for the sake of it (let me restate the fact that Ultra settings are not and have never been the baseline for actually playing the game at, you can't actually see the details above High unless looking for them in a zoomed in image.)

The fact that skyrim d3d9 of all things reaches nvidia performance levels, with a wrapper that has no client side optimisations to enable such gains and is just taking advantage of the drivers capability to talk with the hardware in a more optimal way is evidence enough that you're wrong.

Worst is you keep returning to games that were demonstrated to have Radeon Driver bugs at the time to make your case.


Heres a funny bit of dis-optimisation in the AMD driver, it supports D3D11 Deferred Contexts, but these have been demonstrated in drawcall tests to reduce performance unless you also support D3D11 Multithreading.

You brainiacs need to stop "vouching for AMD no matter what", They are procreating their own market share by blaming everyone else for their own lack of engineering.

I also need to make it clear, because someone that is architecture ignorant is going to come in as they always do and say "AMD can't do this or that because of their scheduler"

That is BULLCRAP.

The drivers job is taking the API being fed into it and making it as optimal as possible for the hardware using a JIT Compiler, the performance issues in the AMD driver occur at the API<>Driver point, not the JIT compiler and certainly not the hardware.

AMD's driver is multithreaded, it is so under DirectX12, it is so under Vulkan, theres no reason not to take what they learned (Exactly as Nvidia did) from D3D12 and enhance their D3D11 UMD.

a Generic enhancement to D3D11 overall will also reduce their need to optimize games individually, which occasionally breaks things moving between architectures (won't do anything for the majority which launch broken of course).

Kinetica has no Vulkan target, its D3D(previously debug only) or Sony's equivalent of OpenGL.

 

It is a weird thing is that your system is supposed to not being able to enable the SAM feature. I also have the same CPU & board combo (same BIOS version also) alongside the RX5700 but in the drivers I see that the SAM is disabled and cannot be enabled. Would you care to elaborate what's up with your system and how did you make it work?

 

For me, not only did I have to make sure that Above 4G Decoding (you may also have Re-Size BAR Support as well) was enabled but I had to also disable Compatibility Support Module (CSM) but I also had to enable UEFI. Enabling CSM alone would dissble SAM even though everything else was enabled.

 

LOl those last amd driver issues are never going to be fixed it says, they been around for over 2-3 years right? hahaha. ^