Mantle (API)

Mantle was a low-overhead rendering API targeted at 3D video games. AMD originally developed Mantle in cooperation with DICE, starting in 2013. Mantle was designed as an alternative to Direct3D and OpenGL, primarily for use on personal computers. In 2015, Mantle's public development was suspended and in 2019 completely discontinued, as DirectX 12 and the Mantle-derived Vulkan rose in popularity.

Overview

The draw call improvements of Mantle help alleviate cases where the CPU is the bottleneck. The design goals of Mantle are to allow games and applications to utilize the CPUs and GPUs more efficiently, eliminate CPU bottlenecks by reducing API validation overhead and allowing more effective scaling on multiple CPU cores, provide faster draw routines, and allow greater control over the graphics pipeline by eliminating certain aspects of hardware abstraction inherent to both current prevailing graphics APIs OpenGL and Direct3D.

CPU-bound scenarios

With a basic implementation, Mantle was designed to improve performance in scenarios where the CPU is the limiting factor:

Low-overhead validation and processing of API commands;

Explicit command buffer control;

Close to linear performance scaling from reordering command buffers onto multiple CPU cores;

Reduced runtime shader compilation overhead;

AMD claims that Mantle can generate up to 9 times more draw calls per second than comparable APIs by reducing CPU overhead;

Multithreaded parallel CPU rendering support for at least 8 cores.

GPU-bound scenarios

Mantle was also designed to improve situations where high resolutions and "maximum detail" settings are used, although to a somewhat lesser degree, as these settings tax GPU resources in a way that is more difficult to improve at the API level. While Mantle provides some built-in features to improve GPU-bound performance, gains in these cases are largely dependent on how well Mantle features and optimizations are being utilized by the game engine. Some of those features include:

Reduction of command buffers submissions

Explicit control of resource compression, expands and synchronizations

Asynchronous DMA queue for data uploads independent from the graphics engine

Asynchronous compute queue for overlapping of compute and graphics workloads

Data formats optimizations via flexible buffer/image access

Advanced Anti-Aliasing features for MSAA/EQAA optimizations

Native multi-GPU support

Benchmarks

Performance superior to Direct3D 11

Improved performance in Battlefield 4 and up to 319% faster in the Star Swarm demo in single GPU configuration in extremely CPU-limited situations.

Other claims

Easier to port from Mantle to Direct3D 12 than from Direct3D 11 to Direct3D 12

At GDC 14 Oxide Games employee Dan Baker stated that Mantle would address fundamental development challenges that could not be addressed by a retrofit of an existing API. It is hard to optimize for the graphics device driver.