The latest generations PCs are not solely rely on the CPU in data processing. Thanks to collaboration with the GPU can increase your PC performance up to 10 times as much. High-performance PC is usually used to run programs that require complex processing. During this principle applies, so that the PC work faster, buy a CPU (Central Processing Unit) that get results fast. Last few years the situation is slightly different.
Now, the PC has also been strengthened by the processor chip that great performance potential. One of them is prominent GPU (Graphics Processing Unit) that is contained in the graphics card. The latest 3D games require a series of complex processes that must run in realtime. As a result, the GPU was working harder than the CPU. Not surprisingly, graphics cards have also have the 1 GHz processor and super fast memory of large capacity (2 GB).
GPU performance potential of this huge fee that is later also used creative programmers (supported by GPU manufacturers) to perform other complex processes, such as video processing, flow simulation, and prediction market prices. Three years ago, NVIDIA has developed CUDA (Compute Unified Device Architecture) which allows calculation of the program (using syntax similar to C programming language) can be transferred to a specific GPU (GeForce 8000 series began). Rival NVIDIA, AMD gave his support to the standard OpenCL of Khronos Group, which now also supported by NVIDIA.
Both these standards aim to divide the process of program processors that support OpenCL (CPU, GPU, or another processor). Microsoft did not want to miss and integrate the collection of special orders (Direct Compute) to the latest API DirectX 11 for program execution process can also be run in the GPU. CHIP discussion this time will describe the mechanisms and calculations on the GPU any program that can utilize the GPU for more optimal PC performance.
Although the software has been available and producers graphics chip (GPU) has also shown its commitment, the program can take advantage of the GPU is still limited in number. In addition, there are still many who question, how can such a fast GPU? To answer that question, in general can be said that the mechanism of the process units (count) is different CPU and GPU.
In a particular CPU models with 6-core, there is a virtual core features (such as Intel Hyper-Threading) to take advantage of CPU processing capability or the total remaining there in the 8-core CPU. This means, 8 program (or thread) can be executed in parallel. CPU core developed for general duties and is flexible because it can overcome various problems and process the threads are completely different in each core in a single cycle (clock rate).