DYNA JOURNAL ENGINEERING

Key words:

CPU, GPU, CUDA, computación paralela, simulación, mecanismo ferroviario, parallel computing, simulation, railway mechanism

Article type:

NOTA TECNICA / TECHNICAL NOTE

Section:

TECHNICAL NOTE

Many computers incorporate advanced graphics cards with very attractive features for computation of engineering problems that require a large number of operations, functioning as a coprocessor of graphic processes or floating point operations (graphics in video games, simulators or rendering). If the central processing unit (CPU) of a current conventional computer has 4 to 8 process cores, the Graphics Processor Unit (GPU) has hundreds or thousands of process cores. The architecture of the CPUs is oriented to the efficiency of the tasks of the operating system and, mainly, to the sequential processing whereas the GPUs, although they execute more basic tasks, they have an architecture oriented to carry out operations in floating point, handling multiple tasks simultaneously, in parallel