The rapid expansion of NVK is truly remarkable!
The rapid expansion of NVK is truly remarkable!
I've completed some tests (though I can't share them on Peertube) and OW2 achieves over 100 FPS in certain areas. This update comes without the new MR, which implements Vulkan 1.1 properly. After merging, DXVK versions up to 1.10.3 will function out of the box. I want to highlight that the initial announcement about NVK was made last March. Eight months later, the improvements are impressive. Next year we might even run full FOSS Nvidia drivers on Linux with FSR 3.0 for upscaling. Plus, everything else runs smoothly on Nouveau+NVK—from Plasma 6 to Sleep and VRR. The FOSS team has made remarkable progress quickly. The only issue so far is HDR in Plasma 6, but the developers will likely address it soon. I'm really excited; there are so many positive developments in Light Speed that I struggle to keep up! 
NVK refers to a free userland for NVIDIA Vulkan drivers running on Linux. DXVK acts as a middleware layer that builds on Vulkan, while the new kernel and NVIDIA driver provide the underlying support.
I own a 4090M and most games freeze, including OW, but with this new MR I'm really optimistic things will improve. Still I managed to run a full FF II stream on DXVK without any issues for about two hours. Playing a bit more today.
Discovered your videos recently; there were several stutters, but it looks promising for this new software. Excited about the potential to use it on more GPUs too.
Overwatch 2 merge proposal (generally acceptable for most users here) This is related to gitlab/dev tools, so it's not widely known outside that context. AMD's DLSS feature focuses on variable refresh rate. As a gaming community hub, these points are fairly standard. No need to clarify further—this is just a reference to NVIDIA and Vulkan names. https://www.collabora.com/news-and-blog/...g-nvk.html https://github.com/doitsujin/dxvk
I'm a software developer too, and I wasn't aware that MR stands for "merge request." We typically use PR (Pull Request) instead, which is what GitHub refers to. This could be partly because we work in an industry where MR means "Make Ready" – the time and effort needed to prepare a machine. It might get confusing if we exchanged messages about MR related to a task involving "Make Ready."
Sure, I get what you mean. MR is quite GitLab-focused, so if you haven’t used it yet, it’s reasonable to become familiar with it now. It actually makes more sense in this context than PR, since you’re merging rather than pulling.