A past attempt resulted in slower code, but that could have been implementation-specific.

There's also a lack of volunteers to maintain that part of the code, so it will be removed from 5.10 for being perpetually broken and hopelessly out of date.

are you referring to the pp tool ?
But what I don't understand is that when the compiler already makes byte-code, how hard can it be to save that result into a file!

are you referring to the pp tool ?

I assuming you're talking about PAR? Then: no. This project was much older.

But what I don't understand is that when the compiler already makes byte-code, how hard can it be to save that result into a file!

Not hard. But from what I gathered, the resulting bytecode was huge and allegedly it took a longer time to load this image from file, than to compile it again from scratch.

No references to point to (not at this time anyway), sorry. It's all hearsay and (possibly erroneous) recollections.

update: I found a reference: This Week on p5p 2000/06/18: More Attempts to Make B::Bytecode Faster

The whole point of B::Bytecode is to speed up the startup time of Perl programs. Two weeks ago Benjamin Stuhl reported that bytecoded files are actually slower than regular source files, probably because the bytecoded files are so big that it takes a lot of time to read them in.