Testing some singleton implementations

[Stanislav Blinov]

On Sunday, 9 February 2014 at 12:20:54 UTC, luka8088 wrote:

> What about swapping function pointer so the check is done only 
> once per
> thread? (Thread is tldr so I am sorry if someone already 
> suggested this)

That is an interesting idea indeed, though it seems to be faster 
only for dmd. I haven't studied the assembly yet, but with LDC I 
don't see any noticeable difference between SyncSingleton and 
FunctionPointerSingleton.