30 Jan
2007
30 Jan
'07
2:31 p.m.
Hi, Richard and Paul
I believe that this is where we can gain the greatest speed improvement. If the screen is refreshed often, there is a lot of work for windows to do here and that would cause a delay. <<
On my machine, DOS Fractint completes the default Mandelbrot at 1600x1200 with symmetry=none and passes=1 in 3.16 seconds. In Windows, it's 110.75 sec. That's about 35 times slower! I hope you can get at least some improvement before releasing a general test version. It will make further testing go much faster! Then we can test arbitrary precision and some slower types, such as the MandelbrotMix4 and MandelbrotBC2 formula types that Jim Muth uses a lot. Lee