Tweet
As anyone who ever reads Phoronix would have seen, sometimes benchmarks show worse performance for Vulkan than OpenGL. Sometimes Vulkan benchmarks also appear to be CPU bound even when they are better than OpenGL results.
So what aspect of Vulkan needs to be further developed? What isn't ready yet?
Is it the drivers that aren't mature enough yet?
Is the API itself in need of a major update? I doubt that's the problem.
Is it because the individual games need to be better optimised for Vulkan? Is the amount of work it takes to make each individual game Vulkan-compatible so high that even the biggest-budget games are having trouble using Vulkan to it's potential?
Or is the immaturity of Vulkan support not on a per-game basis, but due to a shortage a programmers who know how to use Vulkan? Is this the major hurdle that's preventing Vulkan from being widespread while it's young?
I wonder if getting over the current problems would cause Vulkan to gain some adoption over D3D12. I guess that performance on Linux won't motivate developers to support it if it's actually slower than OpenGL on Linux. However, I still think the bigger factors for dev's choosing DX12 are ignorance, stupidity, and brand-recognition for DirectX. Interestingly, these benchmarks found Vulkan on Windows to have lower average driver overhead than DX12. It was significantly lower on both Nvidia cards tested, and barely higher on Radeon (which is surpising given Vulkan's Mantle roots).
Maybe Vulkan would have better adoption if it was a mid-level API like Apple's Metal. Hopefully there will be an API like that coming which can benefit from Vulkan's development. I feel like such an API would be more mainstream due to a lower barrier of entry.
So what aspect of Vulkan needs to be further developed? What isn't ready yet?
Is it the drivers that aren't mature enough yet?
Is the API itself in need of a major update? I doubt that's the problem.
Is it because the individual games need to be better optimised for Vulkan? Is the amount of work it takes to make each individual game Vulkan-compatible so high that even the biggest-budget games are having trouble using Vulkan to it's potential?
Or is the immaturity of Vulkan support not on a per-game basis, but due to a shortage a programmers who know how to use Vulkan? Is this the major hurdle that's preventing Vulkan from being widespread while it's young?
I wonder if getting over the current problems would cause Vulkan to gain some adoption over D3D12. I guess that performance on Linux won't motivate developers to support it if it's actually slower than OpenGL on Linux. However, I still think the bigger factors for dev's choosing DX12 are ignorance, stupidity, and brand-recognition for DirectX. Interestingly, these benchmarks found Vulkan on Windows to have lower average driver overhead than DX12. It was significantly lower on both Nvidia cards tested, and barely higher on Radeon (which is surpising given Vulkan's Mantle roots).
Maybe Vulkan would have better adoption if it was a mid-level API like Apple's Metal. Hopefully there will be an API like that coming which can benefit from Vulkan's development. I feel like such an API would be more mainstream due to a lower barrier of entry.
Comment