nVidia Optimus

The integration of PCI Express lanes to the processor and the launch of Intel P55, nVidia took out of the market of chipsets for Intel, focusing on the development of plaques dedicated. Thus, we have a situation where most users use a PC with an integrated Intel chipset and they need to perform better install a dedicated GPU. This also applies to notebooks, with the difference that the choice between a model with or without a dedicated GPU is made at the time of purchase.

Another problem in the case of notebooks is the issue of power consumption, as a dedicated board consumes much more energy than included in the IGP chipset, which is a shorter battery charger. To increase sales, nVidia started to develop hybrid systems that allow switching between the two, in a saga that can be divided into three chapters.

The first generation came to market in 2006 and was used in models like the Asus N10JC and Sony Vaio SZ-110B. In them, switching was still done manually through a switch, and you need to restart each time you wanted to change. Thus, the majority had preferred to use one or the other all the time, keyed rarely, if ever.

The second generation can be found in models made from 2008, where the switching is now done via software. In them, it is still necessary to make the change manually (by changing the option in the properties of energy), but but at least no longer necessary to restart. The big problem was that the switch is far from clear (the video was blocked by 6 to 12 seconds) and you must close all applications that used functions or 3D acceleration (games, video players using the decoding in hardware, etc..) before switching.

The Optimus is the third generation technology, announced in February 2010. The advantage for the two previous generations is the fact that it is based on an automatic switching system, capable of switching between the video chipset integrated GPU and dedicated, according to the applications. Thus, the integrated chipset (which offers a lower power consumption) is used tasks of the day, while the GPU kicks in when running games and other tasks that require 3D acceleration and video decoding in hardware.

Instead of being used separate outputs for the IGP and GPU with hardware-switchers (which increased costs and made the slow transition), the Orange works in a much more elegant way, using the PCI Express bus for data transmission frame-buffer (technology that nVidia christened "Optimus Copy Engine").

When the GPU is idle, the IGP does all the work, with updates from the frame-buffer is copied to a memory card dedicated. Because the data is always current, you can make the switch quickly. There is still a little lag transition (from 200 to 300 ms), but it will pass unnoticed, as it enters the load time of application.

Since the GPU is enabled, the IGP enters a low-power mode, where only the frame-buffer and the conversion circuits are active video. Thereafter, the GPU is to do all the processing, copying the frames already rendered to the frame-buffer of the IGP, which now operates only as a kind of printer, sending images to the monitor.

esides reducing the transition time, the new system has reduced costs, because instead of two separate outputs connected to the monitor for switchers, it was used a single output, connected to the IGP. The GPU has to behave as an "accelerator" in the strict sense of the word.

Activation of the GPU is based on the application. The drivers include a list of applications validated, providing a series of profiles, which can also be adjusted manually. The list includes not only games but also applications where the functions of acceleration, encoding and decoding video GPU results in performance gains. There are also cases where the GPU is used to reduce electrical consumption, as in the case of applications with support for CUDA or OpenCL, where the GPU can do the job with less power consumption.

The profiles can be configured through the display settings, which allows you to subscribe to the settings for specific applications. You can also manually choose between using the IGP or GPU clicking the right button on the icon of the application, instead of opening it directly.

The idea of nVidia is making the system more attractive and facilitate the development of notebook and netbook hybrid, allowing manufacturers to implement the system by simply adding a MXM card with the GPU on the existing platform. It's similar to what was done in the ION (which will offer enhanced graphics in netbooks with the Atom), but with a much wider application.

As always, the use of dedicated GPU results in increased cost, but at least the difference fell in previous generations, making the technology more palatable for those who want to use the notebook to play sporadically. Another important point is that the removal of switchers and additional tracks and a reduction in electrical consumption have opened the door for use in portable notebooks, creating a new class of laptops for casual gaming. A good example is the ASUS N82JV, which combines a GeForce G220M Core i5 and a notebook with screens 14 "and 2.3 kg: