When Mantle was demoed in 2014, the improvements found on low-end CPUs were quite shocking freeing up the CPU to do other things, or indeed, pipe more instructions to the GPU if it could cope. Oddly, this doesn’t necessarily increase the overall performance of the GPU directly (however, Vulkan does include a number of improvements in that regard), the biggest advantage comes in the form of low overhead. Removal of this code has been brought up many times over the years, for each progressive release of OpenGL.Ī new API releases OpenGL from these archaic shackles, as well as introduce almost direct access to the hardware on modern graphics processors, sometimes referred to as bare-metal programming. Over the years, a lot of ‘cruft’ has found its way into the code, leading to poorer performance on modern hardware. One of the long standing issues with OpenGL over the years is its requirement for backwards compatibility for specific hardware and software. Vulkan is not a continuation or renaming of OpenGL, but a new and separate API.
Much of this work helped with the development of the new API that later became known as Vulkan. Rather than simply cease development of Mantle, AMD submitted its ideas and work to Khronos Group, the maintainers of the venerable OpenGL API.
