Замонавий график процессорлар архитектураси ва технологиялари
Аннотация
В данной статье анализируется архитектура современных графических процессоров и основные особенности выполнения программ в GPU, анализ платформ программирования используемых в графических процессорах, а также методы оптимизации программного кода.
Маколада замонавий график процессорлар архитектураси тахлили ва уларда дастур бажарилишининг узига хусусиятлари, график процессорларда кулланиладиган дастурлаш платформаларининг тахлили, дастур кодини оптимизациялаш усуллари келтирилади.
The widespread need to implement high-quality, interactive 3D graphics has led in recent years to the significant technological development of graphics processors (GPU)s, which are an integral part of any personal computer. Graphic the processor has acquired the quality of a high-performance device based on the use of parallel technologies. At the same time, the graphical processor provides the ability to program the processing of initial data at the level of direct commands of the graphic processor. High-quality 3D graphics, which requires parallel processing of resource-intensive data from the GPU, led to the creation of a special, in a certain sense, unique GPU architecture. The modern graphics processor is designed so that the majority of its transistors in the processor chip are used for data processing, to the detriment of caching and control of the progress of the program.This article analyzes the architecture of modern graphics processors (GPUs) and the main particular qualities of executing programs in the GPU. The possibility of using the GPU for solving resource-intensive numerical problems such as the basic algorithms of digital signal processing and images is considered. The problems and limitations that arise when computing from the central processing unit to the graphic processor are transferred, the analysis of the programming platforms used in the GPU is analyzed, and variants of optimizations of applications using the graphics processor are also offered.
Маколада замонавий график процессорлар архитектураси тахлили ва уларда дастур бажарилишининг узига хусусиятлари, график процессорларда кулланиладиган дастурлаш платформаларининг тахлили, дастур кодини оптимизациялаш усуллари келтирилади.
The widespread need to implement high-quality, interactive 3D graphics has led in recent years to the significant technological development of graphics processors (GPU)s, which are an integral part of any personal computer. Graphic the processor has acquired the quality of a high-performance device based on the use of parallel technologies. At the same time, the graphical processor provides the ability to program the processing of initial data at the level of direct commands of the graphic processor. High-quality 3D graphics, which requires parallel processing of resource-intensive data from the GPU, led to the creation of a special, in a certain sense, unique GPU architecture. The modern graphics processor is designed so that the majority of its transistors in the processor chip are used for data processing, to the detriment of caching and control of the progress of the program.This article analyzes the architecture of modern graphics processors (GPUs) and the main particular qualities of executing programs in the GPU. The possibility of using the GPU for solving resource-intensive numerical problems such as the basic algorithms of digital signal processing and images is considered. The problems and limitations that arise when computing from the central processing unit to the graphic processor are transferred, the analysis of the programming platforms used in the GPU is analyzed, and variants of optimizations of applications using the graphics processor are also offered.