ARM, Mali and making it real

OpenCL on ARM-based systems

OpenCL implements a control-slave architecture, where the host processor (on which the application runs) offloads work to a computing resource. When a kernel is submitted for execution by the host, an index space is defined. The index space represents the set of data that the kernel will be applied to. It can have 1, 2 or 3 dimension (hence the name of NDRange, or N-dimensional range). The instance of a kernel executing on an individual entry in the index space takes the name of work-item. Work items can be grouped into work-groups, which will execute on a single compute unit.

Kernels can be compiled ahead of time and stored in the application as binaries, or JIT-compiled on the device, in which case the kernel code will be embedded in the application as source (or a suitable intermediate representation). The kernel can be compiled to execute on any of the supported devices in the platform.

The application developer defines a context of execution, which is the environment the OpenCL C kernels execute in. The context includes the list of target devices, associated command queues, the memory accessible by the devices and its properties. Using the API, the application can queue commands such as: execution of kernel objects, moving of memory between host and processing plane, synchronization to enforce ordered execution between commands, events to be triggered or waited upon, and execution barriers.

The architecture of the Renderscript API is analogous to OpenCL.

OpenCL enables general purpose computing to be carried out on the GPU. The ARM Mali-T600 series of GPUs has been specifically designed for general purpose GPU computing, and an OpenCL 1.1. Full Profile DDK as well as a Renderscript DDK is available from ARM.

At this year’s ARM Technology Conference in Santa Clara, I will present two pieces on this topic. My first presentation will focus on OpenCL and how this is enabled on ARM based systems. I will discuss key aspects of the OpenCL architecture and how the API is used, as well as highlights of the OpenCL C programming language through some example code.

In my second presentation I will dive into more detail on GPU Computing to explore some practical use cases and how the design considerations of the ARM Mali-T600 series of GPUs make them the perfect fit for compute frameworks such as OpenCL and Android Renderscript.

The aim of these sessions is to help you to understand the applicability of GPU Computing and how you can get started and explore it yourself.