QuoteIt took a while, but here finally a V1.8 update (http://users.skynet.be/fquake/MandelDX11.zip) enabling double precision calculations for really deep zooming.
It was waiting for the June DX SDK to get a HLSL DX11 compiler that can handle doubles in loops... After that a straight port of my float versions did not result in a working renderer with doubles on ATI GPUs. There seems a bug in the drivers that prevents normal calculation.
After some experimentation I could get a doubles version '2' that works on ATI. By default it will use the intended doubles version '1' that happens to run fine on Nvdia GPUs. It is possible to toggle between the two versions as is described in the release note.
[via] (http://forum.beyond3d.com/showpost.php?p=1468115&postcount=85)
I made a video (http://player.vimeo.com/video/14678003?byline=0&portrait=0)