The latest update to 3DMark adds Vulkan support to its API Overhead test, which attempts to render as many simple objects as possible while keeping above 30 FPS. This branch of game performance allows developers to add more objects into a scene, and design these art assets in a more simple, straight-forward way. This is, now, one of the first tests that can directly compare DirectX 12 and Vulkan, which we expect to be roughly equivalent, but we couldn’t tell for sure.
While I wasn’t able to run the tests myself, Luca Rocchi of Ocaholic gave it a shot on their Core i7-5820K and GTX 980. Apparently, Vulkan was just under 10% faster than DirectX 12 in their results, reaching 22.6 million draw calls in Vulkan, but 20.6 million in DirectX 12. Again, this is one test, done by a third-party, for a single system, and a single GPU driver, on a single 3D engine, and one that is designed to stress a specific portion of the API at that; take it with a grain of salt. Still, this suggests that Vulkan can keep pace with the slightly-older DirectX 12 API, and maybe even beat it.
This update also removed Mantle support. I just thought I’d mention that.
Scott, does this mean
Scott, does this mean equivalent gaming will be achievable on linux in the not to distant future? I am sure there is much more to it, but I so hope we are finally getting close.
I’m a bit confused over the
I’m a bit confused over the connection here. How do you think futuremark adding a benchmark to to a windows only program translate to better performance under linux.
Because Linux is congregating
Because Linux is congregating around Vulkan, concentrating optimization efforts.
As for the question? Eh. Like I said, it's a single test on a single system on a single, limited-focus application. Not much to draw conclusions from, except that it's got the potential to be competitive. I wouldn't think too much more about it yet.
copy that
thanks scott
copy that
thanks scott
Anyone that got the full
Anyone that got the full benchmark suite tested dx12 vs mantle ?
how did mantle compare ? (before it was removed)
mainly curious to see if vulkan (based on mantle) got better or worse
Here is the funny part with
Here is the funny part with the latest Futuremark benchmarks.
Time Spy was the only maybe benchmark where new Nvidia cards where getting performance improvement from async. Right?
The Vulkan benchmark is probably the only Vulkan software code that seems to run better on Nvidia cards than AMD cards.
Yes, yes I know, I have an AMD logo as an avatar, so I am probably wrong and blah blah blah.
Maybe nVidia hardware is
Maybe nVidia hardware is better utilized to begin with, and as such don’t gain as much from async as underutilized AMD hardware.
AMD’s hardwware is
AMD’s hardwware is underutilized because of Games not being written to take advantage of the Vulkan API and async-compute, and it will take time but AMD’s hardware, even the older hardware will improve over a longer time. Even Nvidia’s hardware shows some improvment under Vulkan but Nvidia’s older hardware will not improve as much relatively as AMD’s older GPU SKUs(GCN 3/”1.2″ and GCN 2/”1.1″) and that’s just fine by JHH as he wants more Nvidia new hardware sales.
Let’s wait for PCPer to test some Polaris Refresh 500 series SKUs and on more Fully Vulkan API enabled titles to see how things have improved for AMD Mainstream SKUs, as any Flagship testing will have to wait for Vega to arrive.
Also with respect to any testing of Fully Vulkan Enabled Games Lets test some Vulkan API managed dual GTX 1060s against a Dual RX 580s setups with both of the Nvidia and AMD respective dual GPU confgurations using Vulkan’s API managed GPU mulit-adaptor with no CF/SLI(Not support on the 1060 anyways) dual GPU allowed.
I think that Vulkan’s API managed GPU multi-adaptor will be great for GTX 1060 and 1050 owners that mwy want to use dual configurations and the Vulkan API will be great for AMD’s GCN GPUs as the games/gaming engines begin to make use of the Vulkan APIs full feature sets and async-compute.
Windows 10(DX12 by extention) is not an option for some gamers and hopefully there will be more windows 7, 8.1 OS and Vulkan API testing done as well as the Linux/Vulkan and Steam OS/other linux distro gaming testing under the Vulkan API, because Vulkan is truely a cross platform graphics API.
GoW 4 also profits from async
GoW 4 also profits from async on vs off on Pascal Geforce cards, though less so (2-3%) than Radeons (7-9%) – which one could chalk up to Nvidia’s hardware being utilized better already even with async off.
Before we go down the road of
Before we go down the road of “DirectX is dead”, let me remind everyone:
Optimized OpenGL > DirectX > OpenGL
Same effect here:
Optimized Vulkan > DirectX > Vulkan
DirectX is significantly easier for developers to work with, in large part due to better tools being available for use.
The Best Developers’ tools
The Best Developers’ tools comes with the largest install base, so with Vulkan being used across the non-Apple mobile markets, in addition to PC/Laptop market, expect that the tools, Middleware, and other features will be available to help the poor script kiddies cope with any Vulkan API usage.
Really already the Vulkan hard stuff has tools in the SDKs to help with automating things for the “Developers’ in gaming. And any real Developer that can not get up to speed with any API is maybe just not be cut out for a developer job in the first place. Also Vulkan on mobile is the same Vulkan that is used for the desktop, so there will be plenty of trained Vulkan programmers across a much larger market than the relatively small market represented by DX12 and its official market of only under M$’s windows 10 thumb.