Thought though.. Does the comparing each line of screen data to see if it has changed, add a bit of extra cpu usage on the host side?
Just thinking that if it does, older machines might suffer if they are already devoting a lot of CPU to the game itself. Perhaps a full toggle is need to use old\new way?
Well, I wouldn't say the comparison adds extra CPU usage (though technically it does, and in the very unlikely absolute worse case scenario it could be theoretically noticed), but instead, logically, it mostly moves the CPU usage from one place to other relieving the scalers from its heavy workload, helping to make them do the work that's really only needed. I sincerely don't believe that old computers will suffer at all, but on the contrary those are most likely the ones that would probably notice a bigger improvement. I believe the optimized scalers only have one drawback, and that is the usage of about 2MB of extra memory, which is really not that much anyway.
By the way, on the old/new issue, you can always compile the "original" scalers in DOSBox by defining UNOPTIMIZED_SCALERS during compilation.
Please make sure the first post (didn't check but I dont remember it being updated) is updated with the latest build so new comers wont have to sift through all the messages trying to find the latest exe of your patch 😀
I've never put the exe in the first post... only the diff patches. (this is in part due the limitation of the posts to have no more than 5 attachments)
Perhaps I should only put the latest patch and the latest binary there...
"Time is of the essence"