In practical terms though, the assistance of the integrated graphics core is hardly going to add much to the performance of the discrete part, is it? If the discrete card wasn't already much more powerful than the integrated chip then there would be no point in implementing this solution in the first place. On the up side, nVidia already has SLI ready to go so it's simpler and more seamless for them to use this approach rather than designing some kind of switch to control which adapter has the output duties at any given time.

will you may read that article again, what you said is true, but thats only if more power is needed, or MAximum performance, but when the power save mode, the dicrete GPU will be turned off to let the IGP handle the job...

I know, they just added the Higher performance point where both GPU's will work together in SLI, but the point is the same...

Nothing you have to do as a end-user. If you are not running games, IGPU is running. If you are running any game or application that requires rendering power, discrete GPU kicks in. You don't really see what's happening except for the performance.