[BigETI]

Quote:

Originally Posted by ****** That benchmarking code, despite your claims to the contrary, is NOT valid. You are comparing 1000001 dialogs (yes, not 1000000, you don't even know how many tests you're running) to 1 dialog - of COURSE your system will be faster! "CallLocalFunction" does NOT call functions INSTANTLY - nothing happens INSTANTLY, it just happens very quickly because there is only a couple of public functions to search through! If you wanted a true VALID test, you would have to make 1000001 "DIAL" callbacks, put them all in your mode, compile, then run that test to see the difference. Check the benchmarking code I used to test y_commands. I wrote a file with about 676 command skeletons, then used macros to repeatedly call that file - once generating y_commands code, once generating ZCMD code and once generating strval code. This way I KNEW that all the systems had the same test parameters and I could go experimenting from there. Make sure you also choose a good sample of dialogs to test call - there's no point always calling the first one or the last one, you need to call a large number well distributed through the available IDs for EVERY test. As it is this is complete rubbish!

Ok I got what you've posted.
I'll create myself thousands of callbacks/checks in a file and paste it in the benchmark script to create a proper benshmark test. I'll post the results and scripts of the if else, switch, B-DIAL, FDLG etc. benchmark tests