Only suggestion I might have re speeding things up is that IIRC, it's faster for a DOS program like Fractint to access Extended or DPMI memory than the older Expanded memory standard. So if your OS (such as OS/2) provides the option to control which type of memory it makes available to Fractint, giving it a chunk of Extended or DPMI memory might make it go a bit faster. On 2 Jan 02 at 10:44, Bill Jemison wrote:
<<...do you have any tips for optimizing rendering time and improving <<performance re: the cache settings?
No. But I always thought that whether one is writing to screen, memory or disk, that the limiting factor was the actual calculation time, not the write time. But that is just my conjecture, as I know nothing about the reality of the situation.
BTW, I've never used background generation for its speed - seems to be about 25% slower than screen generation for me...but it is handy if I've got something else to do <g>
Under most multitaking OSes, background processing will be slower than screen generation - when running in full screen mode, Fractint would be getting much more of the processor time than when running in a window in the background. I run it regularly under OS/2, generating to memory, because I *do* like to use my computer for other things at the same time, and Windows95 and W98 just can't do it without erratic response and momentary freezes in the foreground Windows app. David gnome@hawaii.rr.com