Thanks for the hints to everyone here
FOV: changed it to 55. It's superior to the default ratio.
Multiview: sorry for the confusion due to the posting of mine, Joe. When I wrote "I experimented without the multi-view setting and discarded it very quickly, because the image became distorted and the rfdynhud info was spread to the left and right edge of the left and right display." I actually meant I discarded the non-multiview setting. That means, multiview was and is active most of the time.
Bezel adjustment: thrown out for now. I know, in a real car the front column takes away some visibility when looking to the left, but in a formula car I don't want that. and in the GT2 Porsche I also prefer this setting, as I see no reason to properly adjust FOV and bezel adjustment that the Porsches front column aligns with the bezels on the left hand side of the middle display.
Frames: good in F1 1985 with all cars on track. Lowest fps rate at the start (I deliberately started from the end) with 23. Apart from that, very stable in the low 30s. For the SCES LeMans race this area will need some work. When the ferris wheel is in sight frames drop to around 16, regardless of the daytime and how many cars are on track. Furthermore the genuine test was performed in the night and the frames dropped below 20 at too many occasions. It was manageable, but I consider it too risky. 23 should be the acceptable minimum.
EDIT: Oooops... I just noted I still had these values in my plr-file:
Max Headlights="12" // Max headlights visible relative to your car.
Headlights On Cars="1" // Headlights illuminate other cars.
I changed that to 3 and 0, as pointed out by Doug here: http://simhq.com/forum/ubbthreads.php/topics/3296536/Event_3_Le_Mans_Night_Driving_.html#Post3296536
But if the system still had barely manageable frame rates than with all cars in sight I am optimistic