Cycles render in Blender for a very long time has been supporting GPU compile, GPU Compile option is like a switch it lets you switch to your dedicated Graphics Card if your system has one for rendering. GPU Compile can be faster than CPU rendering, depending on your system configuration, this means it would take lesser time to render and the rendering task is done by a separate hardware optimised for rendering, this can also help reduce the viewport lag, because the CPU is not in use and can let you do minor changes faster in real time.
GPU Compile has been my long time favorite and was always set to default in my projects while setting up rendering, which at that time only supported Nvidia Graphics Cards that supported CUDA. Until I had switched to a MacBook Pro 15 2015 (with dedicated graphics), which came with an AMD Radeon R9 M370X. To enable GPU compile for the AMD graphics card on the mac you needed to use OpenCL test flags, until release 27.5.
You can learn how to enable GPU compile for AMD graphics card over here. Blender 2.76a came with support for AMD Graphics card, that used OpenCL and supported the following Graphics cards, this meant, I could finally use GPU Compile and render my projects with the Graphics Card, this was not very stable and had issues like loading of the rendering Kernel and other strange bugs, until Blender 2.76b which improved the stability by a larger number and now would render most of my scenes with crashing. It still needs a lot of improvement until it could be used in production on a large scale, but it is nice to see support for AMD Graphics card, this lets users have a wider choice of Graphics card option and support.
I still use my Alienware M17 for rendering, which comes with Nvidia Graphics card (Nvidia GTX 580M). Thank you for rendering, If you have any question feel free to leave a comment below.