Testing the Nvidia Optimus Technology – Operation

As we mentioned in the introduction, with Optimus, NVIDIA aims to provide the user a solution to benefit both 3D performance to match its needs without neglecting energy consumption.

To meet this challenge, the solution is simple. Just implement two separate graphics circuits, with one side an integrated chipset and the other a dedicated graphics card. Nothing new in fact, many laptops are already equipped with this system for many years.

But until now, the user had to manually switch from integrated graphics to the dedicated graphics card or vice versa. If the arrival of Windows Vista has simplified the operation by allowing fail-over “hot”, that is to say, without having to restart the operating system, the constraints were too great for many users.

Several studies by NVIDIA have indeed shown that most people with a laptop with a dual graphics system, were content to use only one of the two. The reasons are many, but most often, users simply did not know of this possibility of switch as needed. For others, having to press a button and wait a few seconds is just too restrictive.

In the end, only a small percentage of users actually exploit the opportunities offered by the presence of a dual chipset. NVIDIA decided to respond and develop a fully transparent system for the user to automatically switch from integrated graphics chipset dedicated as needed.

If you regularly follow the High-Tech news, you are certainly aware of the peculiarity of the latest Intel Arrandale processors, better known by the appellation Core i. They have indeed a new architecture that includes in a single chip, the processor (CPU) and an integrated graphics (IGP). The latter most often coexists with a dedicated graphics card (GPU) to provide 3D performance more suited to the needs, especially those of the players. Only problem so far, few models offered the possibility to the user to choose between the IGP and GPU.

The Optimus technology developed by NVIDIA thus aims to overcome this limitation by allowing the automatic processing of the integrated graphics card and chipset. No need therefore to perform this task manually.

In practice

NVIDIA likes to compare the new technology with the operation of a hybrid vehicle. The user must have continuously the power needed but it must be exploited wisely. It therefore should not be wasted in order not to cause unnecessary energy consumption and penalize autonomy.

In developing its new technology, NVIDIA has simply inspired from the past. Previously, the user had to anticipate his needs in terms of power and activate 3D calculations, if any graphics card before starting the program concerned. For example, one game.

With Optimus, the preliminary step is no longer necessary. The graphics driver is able to choose himself and in real time between the IGP or GPU. To do this, it is based on a series of pre-recorded profiles, regularly updated by various software vendors, as verified by NVIDIA. For example, if you write text in Word, the driver knows that this type of program does not require significant computing power. So glad the integrated chipset. However by launching Call of Duty, this time the driver will recognize that this is a video game and activate immediately, the dedicated graphics card. The latter being more appropriate.

During our test period, we did not encounter any particular problems. The automatic switch is, as expected, really instant (200 ms announced). And if NVIDIA was not available to us a little widget to visualize in real time the active chipset, it would have been impossible to notice the passage of the IGP to GPU.

Automatic management has many advantages, it is undeniable. At the point to ask very quickly why this simple system has not been proposed earlier. Still, as any automated system, sometimes the Optimus driver from being “trapped” by enabling the chipset not always the most appropriate. Do not worry, it is of course possible to regain control.

In this case, simply visit the driver parameters and set manually if for the relevant program you want to use integrated graphics or on the contrary, the GPU. Example type games. If the majority of them, the dedicated graphics card is undoubtedly the most suitable for some older titles just greedy resource 3D, the GMA HD chipset will prove quite up to par.

It is also possible to occasionally choose a chipset by right-clicking the icon of the relevant program before launching it.

Currently, Optimus drivers are specific to models with this technology. However, in the coming weeks NVIDIA will simplify the situation by offering only one more driver for all its mobile graphics cards.


Close this section by a small technical point on the architecture used.

Previously when two graphics circuits were established, it was necessary to use a multiplexer system to cover all the possibilities of displays. Besides latency to move from one multiplexer to another, this complex system of circuit integration involves additional costs for manufacturers.

With Optimus, NVIDIA has revised its implementation. As you can see in the diagram above, now it is no longer necessary to use multiplexers. Why? Simply because the dedicated graphics card is active or not, the display is always given to the video output of the integrated chipset (aka PGI). It is no longer necessary to double the circuits for the GPU.

But then how to ensure optimal performance without being limited by those of the IGP? To get NVIDIA uses a hardware component, the Copy Engine. The latter is responsible for reinjecting asynchronously the data generated by the graphics card dedicated to video controller IGP via a PCI-Express line. The integrated chipset then no longer load calculations when the GPU is activated.The Optimus technology is therefore based on both the software part but also on a particular architecture described above. Therefore, even if your laptop is equipped with a compatible graphics card (GeForce 200 and higher), you can not expect to benefit from a portable non Optimus originally equipped by the manufacturer.