Benchmark specialist Futuremark has announced what is claimed to be the world's first independent test designed to quantify the overhead of different APIs, starting with DirectX 12 and AMD's Mantle.
When AMD launched Mantle in September 2013, it did so with the claim that modern application programming interfaces (APIs) like DirectX and OpenGL carry a major overhead that can reduce performance. This, the company claimed, was particularly noticeable when pairing an otherwise powerful graphics card with a below-average processor - such as with the company's accelerated processing unit (APU) products. Mantle as a standalone API may be no more, but AMD's message has been heard: both Microsoft's DirectX 12 and the Khronos Group's Vulkan APIs have been created specifically to reduce the overhead of their respective predecessors, with AMD's input.
Naturally, claims of lower overheads need proof - and that's where Futuremark comes in. The latest update to its popular 3DMark gaming benchmark introduces the API Overhead Feature Test, claimed to be the first independent method for measuring the differences in API overhead. Initially supporting DirectX 11 as a baseline and DirectX 12 and Mantle as overhead-reduced improvements, the test is near-unique in being more about comparing the performance of different APIs running on one system than different hardware running under the same APIs.
The test works, Futuremark has explained, by making a steadily increasing number of draw calls through each API. The result given at the end is the maximum number of draw calls per second achieved by the API before the frame rate of the rendered scene dropped below 30 frames per second. This test can then be repeated on the same hardware for each API, offering an insight into the improvements on offer.
The DirectX 11 API Overhead Feature Test is available on any system with graphics hardware offering DirectX 11_0 feature level compatibility and at least 1GB of VRAM, while the Mantle test requires AMD-specific hardware with Mantle API support. Those wanting to compare to DirectX 12 will need the Windows 10 Technical Preview Build 10041 or later. The test is available only to 3DMark Advanced Edition and Professional Edition users, and is locked out on the demo and Basic Edition releases.
More information is available on the official website, while a video demonstrating the test is reproduced below.
When AMD launched Mantle in September 2013, it did so with the claim that modern application programming interfaces (APIs) like DirectX and OpenGL carry a major overhead that can reduce performance. This, the company claimed, was particularly noticeable when pairing an otherwise powerful graphics card with a below-average processor - such as with the company's accelerated processing unit (APU) products. Mantle as a standalone API may be no more, but AMD's message has been heard: both Microsoft's DirectX 12 and the Khronos Group's Vulkan APIs have been created specifically to reduce the overhead of their respective predecessors, with AMD's input.
Naturally, claims of lower overheads need proof - and that's where Futuremark comes in. The latest update to its popular 3DMark gaming benchmark introduces the API Overhead Feature Test, claimed to be the first independent method for measuring the differences in API overhead. Initially supporting DirectX 11 as a baseline and DirectX 12 and Mantle as overhead-reduced improvements, the test is near-unique in being more about comparing the performance of different APIs running on one system than different hardware running under the same APIs.
The test works, Futuremark has explained, by making a steadily increasing number of draw calls through each API. The result given at the end is the maximum number of draw calls per second achieved by the API before the frame rate of the rendered scene dropped below 30 frames per second. This test can then be repeated on the same hardware for each API, offering an insight into the improvements on offer.
The DirectX 11 API Overhead Feature Test is available on any system with graphics hardware offering DirectX 11_0 feature level compatibility and at least 1GB of VRAM, while the Mantle test requires AMD-specific hardware with Mantle API support. Those wanting to compare to DirectX 12 will need the Windows 10 Technical Preview Build 10041 or later. The test is available only to 3DMark Advanced Edition and Professional Edition users, and is locked out on the demo and Basic Edition releases.
More information is available on the official website, while a video demonstrating the test is reproduced below.
MSI MPG Velox 100R Chassis Review
October 14 2021 | 15:04
Want to comment? Please log in.