Dolphin Emulator Drops D3D12 Backend, Focuses On Vulkan

[Electric-Gecko]

Good to be going for Vulkan now. But it seems questionable why this (and countless other programs) would go for D3D12 in the first place. To me it's just obvious that it's not worth bothering with while Vulkan exists..

The D3D12 backend showed the most benefit for Intel iGPU users that could support D3D12, but not Vulkan. This was a lot of users. Intel iGPUs dominate our top GPUs used stat in analytics.

Really? D3D12 but not Vulkan on Intel iGPU? I see how that would be a motivation for D3D12, but that's pretty short-term thinking (as they might support Vulkan later). Also, don't both low-level graphics API's have more benefit when you have a fast GPU and slow CPU? But anyway, it's good to be going towards Vulkan now.

I have been playing on Dolphin 5.0 quite a lot recently. It's a good upgrade from 4.0 with low-latency gamepad controls and good audio. But 6.0 (or even 5.1) will be really cutting-edge. Any idea when the next Dolphin will be out?

[Awesomeness]

BEFORE Vulkan was publicly released iirc.

No need doe guess. Just read the excellent Wikipedia article: https://en.wikipedia.org/wiki/Dolphi...2.80.932016.29

What's the point as it can't even couple to Wiimotes without crashing the bluetooth daemon.

It's been a year or so, but I vaguely recall using a Wiimote over bluetooth with Dolphin without issue.

[donkeykong]

My experience with dolphin-vulkan and radv (on a r9 fury): it's a very nice improvement. With some games, with 4k upscaling, radv gives me a solid 60fps where radeonsi ogl is at 37fps. So, vulkan is a performing way better. Only frustrating thing, is that I have a several important glitches (which makes the game completely unsuable) on a few games. I reported the bugs but they don't raise much interest, unfortunately.

https://bugs.freedesktop.org/show_bug.cgi?id=100430

[Hi-Angel]

The D3D12 backend showed the most benefit for Intel iGPU users that could support D3D12, but not Vulkan. This was a lot of users. Intel iGPUs dominate our top GPUs used stat in analytics.

Wow, I thought Windows drivers supported Vulkan long ago. But now that you said I googled around, and turns it's just a few months as Intel supports Vulkan on Windows; more over, whilst GNU/Linux officially supports since 5-th generation, and experimental even earlier, on Windows it's just since 6-th. What a crap, they seems don't even share ANV between platforms.

[Daktyl198]

Seriously? Dolphin has always been a very multi-platform application. Despite this, and people's pleads for Vulkan implementation instead (or at least to be prioritized), they chose to do DX12 anyway. Now they find out the hard way that DX12 wasn't actually the best choice after all. Gee, good job listening to the community. Could've saved a lot of time and effort just doing things the right way the first time around.

A community member is the one who submitted the DX12 backend code in the first place, BEFORE Vulkan was publicly released iirc. They added a Vulkan implementation not too long after its release.

This post is just saying that instead of managing two low-level backends, they've decided to stick with Vulkan and drop DX12.

[sturmen]

Seriously? Dolphin has always been a very multi-platform application. Despite this, and people's pleads for Vulkan implementation instead (or at least to be prioritized), they chose to do DX12 anyway. Now they find out the hard way that DX12 wasn't actually the best choice after all. Gee, good job listening to the community. Could've saved a lot of time and effort just doing things the right way the first time around.

The truth is OSS projects don't work like consumer companies: developers get to work on what's fun for them. DX12 is a perfect example of this: the person who wrote the DX12 backend did it because they wanted to. As I understand it, the lion's share of the work was done by one person who thought it was cool. It wasn't focus-group-tested, it wasn't done by surveying gamers, it wasn't even done by committee. One person did what they thought was fun and it ended up being useful for others. Yet somehow, you got people on forums making a fuss about how they should have worked on other parts of the codebase.

Maintaining code takes effort by other Dolphin contributors. Testers, build automation, triaging issues and answering support questions is time and effort spent by people who may not be so jazzed about DX12. With the disappearance of the prominent DX12 developer, those contributors are freed up to focus on other things. So yes, the developer do not live in a vacuum.

But let me be clear: "listening to the community" is constructive, but never treat open source contributors like they're beholden to you or anyone else's wants.

[e8hffff]

What's the point as it can't even couple to Wiimotes without crashing the bluetooth daemon.

[schmidtbag]

Seriously? Dolphin has always been a very multi-platform application. Despite this, and people's pleads for Vulkan implementation instead (or at least to be prioritized), they chose to do DX12 anyway. Now they find out the hard way that DX12 wasn't actually the best choice after all. Gee, good job listening to the community. Could've saved a lot of time and effort just doing things the right way the first time around.

[sturmen]

No.

In most cases we found the D3D12 backend to be faster. It was faster because it did some threading tricks that the Vulkan backend currently doesn't do. However, that performance advantage slims significantly on higher end systems.

The D3D12 backend showed the most benefit for Intel iGPU users that could support D3D12, but not Vulkan. This was a lot of users. Intel iGPUs dominate our top GPUs used stat in analytics.

While it's unfortunate, we chose to keep the code base easier to maintain, and dropping D3D12 let us do just that. This was the biggest reason. Due to needing a very specific version of the Win10 SDK, retargeting the project to a new version every time the SDK updated wasn't feasible. As it requires buildbot admin coordination as well. And since we have no idea how microsoft intends to manage the Win10 SDK with VS 2017 which we recently migrated to, if microsoft chooses to constantly update it, that's a huge headache for us.

Additionally, the maintainer of the D3D12 backend pretty much disappeared after getting it merged to master, and due to the questionable quality of the code, none of the regular graphics devs wanted to touch it. It fell out of feature parity and users were frustrated that nobody was fixing anything in that backend.

Hopefully the adoption of Vulkan as a first class citizen in prominent projects (including Dolphin) will encourage Intel to support it in future chipsets. Obviously that does nothing to help the computers already sold...