Rub�n wrote:
> The virge acceleration is REALLY cool, in my machine, using very
> optimized 2D rutines, the performance increases using the Virge HW accel
> like this:
> - lines: 20%
> - boxes: 74%
> - hlines: 64%
> - vlines: 13%
It is not very much.... But maybe your main CPU is very
fast. So the speed improvement is not so visible.
Note that these figures should change a lot with the
ping pong buffer techniques. Here, in fact, the speed
gain is very influenced by the rate of ioctl calls
(I guess you don't need any explanation to understand
that boxes drawing take more time than line drawing ;-)
And it seems fbcon handle ioctl rather inefficiently.
Usually, speed gains were over x4 (but usually with
respect to a P150-class CPU) - but on KGI.
Rodolphe
