Switchable graphics refers to mobile computers equipped with both a low-power imbedded graphics chip, and a high-power discreet graphics card, with the ability to switch between them. The imbedded chip does a fine job with everyday applications and video while reserving power, allowing longer battery-driven work periods. The discreet graphics adapter is capable of robust 3D rendering required for gaming, sacrificing power for performance.
A common imbedded graphics chip in many mobile computers is the Intel® Graphics Media Accelerator™ (GMA), also referred to as an integrated graphics processor (IGP). The chip is ideal for everyday use and will save juice when 3D rendering is not required.
A complimentary, discreet graphics card, often a Raedon™ or Nvidia® flavor, makes use of an internal PCI bus and comes with its own on-board memory, drivers and software interface. Such a card is referred to as a graphics processing unit (GPU) and does not overburden the computer processing unit (CPU) with tasks or overhead, or steal system resources like an integrated graphics chip. Instead it carries processing power and RAM on its back, delivering outstanding performance, but requiring more power.
In the past, one had to choose between buying a mobile computer with an imbedded chip or choosing an upgrade to a discreet card, sacrificing either performance or power. This unsatisfactory choice led to the development of switchable graphics with the ability to use the right graphics device for the right job.
The Sony® Vaio™ SZ-110B was the first switchable graphics hybrid to hit the market, featuring both a GMA™ 950 IGP and nVidia® GeForce™ Go 7400 discrete GPU. While a welcomed option, changing from one graphics device to the other requires toggling a hardware switch located above the keyboard, followed by a reboot.
Second generation switchable graphics are handled at the software layer, forgoing the need for a physical switch or reboot. A momentary pause is required, however, while graphic duties migrate from one processor to the other. The handoff is accompanied by some screen flicker, and if an open application is tying up the current processor, it might require closing.
Third generation switchable graphics is transparent, automatic, and virtually instant. Nvidia®’s Optimus™, for example, uses kernel-imbedded processes to seamlessly switch between available graphics devices for optimal performance. The kernel-level library maintains an index of objects and classes along with associations to a particular graphics device. The routing layer manages all open applications, feeding workloads to the IGP or GPU based on the application’s demands. The Optimus™ index is regularly updated, but users can optionally override this behavior and create custom profiles, assigning certain games or applications to the IGP or GPU as desired.
Another example is AMD® switchable graphics technology. Computers with this integrated design automatically employ the low-power IGP when the machine is running on battery power, switching to the discrete Radeon™ GPU when the machine is plugged into an AC outlet. According to AMD® this behavior can be overridden per user preference, and like the former technology, the switch is seamless and virtually instant.
Switchable graphics is an evolving technology with the trend towards automatic handling that does not require user-intervention, designed to avoid the power-hungry or performance-poor trade-off of yesterday. The technology is available in slimline as well as full sized mobile computers.