Re^2: Performance problem with Clone Method

[Commandosupremo]

Dave:

I made the change from the nested maps to a single map and it made a considerable improvement, thank you very much. I've never used the benchmark package, but I think I will start using it as well, its output looks more useful to me than SmallProf's.

[davido]

Glad it helped. By the way, there's a module in the core distribution, Clone, which provides increased flexibility (it can handle datastructures of arbitrary shape) with its clone() function. However, it is actually considerably slower than even your first method. It's an example of where the increased abstraction that makes it a more useful tool all around also carries with it a performance impact.

diotalevi has (or had perhaps) a module on CPAN, Clone::Fast, but when I went to install it with cpan Clone::Fast, the cpan utility couldn't find it. I would like to have tried benchmarking it.

Another aside: Storable (also in the core distribution) offers dclone(), but according to the Clone documentation it's even slower, while being even more flexible.

[Commandosupremo]

I will be sure to give those modules a look over, I don't think they will help me much now (as I am only cloning the matrices), but they may in the future.