Exciting Possibilities as GPUs go Mainstream
GPUs or Graphics Processing Units were designed as “graphics accelerators”. They were a microprocessor, attached to a graphics card, for offloading floating point calculations. It allowed early 2D graphics to run much faster than by using the host CPU directly. The 1990s saw an explosion into 3D graphics with PlayStation and Nintendo 64 as early examples. GPUs were initially used to accelerate the memory intensive work of texture mapping. Later units were added to accelerate geometric calculations such as the rotation and translation of vertices.
In 2002 GPUs added programmable shading – where each pixel could be processed by a short program that included additional image textures as inputs, before it was projected onto the screen. Parallel GPUs had become as flexible as the CPU while they offered several orders of magnitude higher performance than a conventional CPU. This was particularly true in applications requiring massive vector operations.
It was a fact that did not go unnoticed for long. A new stream of development, the GPGPU or General Purpose GPU, has seen the GPU move out of the purely gaming graphics world and into fields such as scientific calculations and image processing, 3D reconstruction and Stock Option pricing determination.
To really go mainstream the programming for GPU’s needs to get easier. The July 22 2010 launch of Nvidia’s CUDA 3.1 Software Development Kit and the Parallel Nsight plug-in for Microsoft’s Visual Studio IDE could be just the ticket…
Loading ...
