The only noticeable speed difference if your scripts are decent is due to unicode and that was not introduced until 6.35. Anyone using older than that is only doing so because they are too lazy to upgrade mIRC, too lazy to upgrade scripts, or too lazy to fix any script problems. Unicode does make mIRC slower because of the nature of unicode. Each character is 2 bytes instead of 1. However, in the majority of scripts (if they are written well), this difference is not noticeable. Regardless of your extremely small and biased group of people, the newest version of mIRC is in use by a very large number of people without a problem.
Just to test speeds, the newest mIRC is a whole 0.3 seconds slower in a basic loop of 100,000 over older version of mIRC. 0.3 seconds? That's not worth noting. And 100,000 loops? Your script(s) shouldn't be looping that many times anyhow in almost all situations if they are written well. And, yes, if you start talking about millions of loops, it will become noticeable. But again, if a script is written well, you shouldn't have to loop anywhere near that many times. If you do, then your script isn't very good.
I run many scripts (23) including socket scripts and many of the scripts are very large (most are part of Invision, which is one of the largest scripts available as far as total number of features available) and there is not any noticeable difference in how the scripts perform. I could double that number of scripts (or more) and still see no noticeable difference.
