Information technology abroad: Internet tablets with quad-core chips to be released by the end of the year

President of Nvidia Jen-Hsun Huang said that his company is about to release quad-core SoC-chips for high-end smart phones and Internet tablets. The president of the company said that first Android-tablets using quad chips will be produced by the end of the year.

The company said that adding more cores to the processor chip is the most appropriate way to improve performance while maintaining cost-effective expenditure of battery power, which is critical for today's smart phones and tablets. To add another core is an alternative to increasing the frequency of the chip. Besides, in the latter case, the processor consumes more power and generates more heat.

According to Nvidia’s forecast, mobile graphical and central processors in the course of their development will imitate the same trend as the processors for PCs and notebooks.

Nvidia is already shipping dual-core GPUs Tegra2 for tablets. These chips operate on the basis of dual-core ARM and GPU GeForce.

In parallel, Nvidia plans to release mobile chips supporting CUDA. CUDA allows to carry out some tasks, such as editing video or working with images, using not only the power of the graphic chip, but also the central chip and vice versa, which greatly speeds up the computation. Until now, simultaneous operation of CPU and GPU chips for accomplishment of resource-intensive tasks has been used primarily in supercomputers.

Nvidia says that graphic chips are now able to perform a limited number of types of calculations, but given the fact that they include more than hundreds of cores, comparing to just a few at the CPU, they are especially good for accomplishing complex tasks related to multimedia, and special mathematical calculations. CUDA combines software and hardware for the implementation of this ligament. Despite the fact that CUDA is used mainly in servers and supercomputers, some desktop programs, such as Adobe Premiere CS5 for video editing, also supports CUDA.