Hi, > Logger is also very fast. While I am sure that you could use a benchmark > program to see a small reduction in overall command line text output > performance, you will not notice it is running. Adding RLE would effect > performance.
I did my own implementation of RLE (feel free to ask for source), and it runs at ~11.5 pentium-clock cycles per Byte(1). This is FAR less then the clockcycles needed to display it on screen, even forgetting about screen scrolling. I doubt RLE would be noticable on any machine. > However, the effect could be improved speed when logging to XMS. > After all, RLE is very quick and would reduce the amount of data that would > need sent to XMS. the time needed to copy to XMS is a constant time to switch forth and back to PM, plus the cost of memcpy. should be fast enough in any reasonable setup. Tom (1) ~12 pentium-clock cycles per Byte for a buffer that consists of a random string of 50% A's, 50% B's. _______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
