in reply to Re: A better way of lookup?
in thread A better way of lookup?

I must admit, like RichardK, I was skeptical about the benefits of using elsifs where the logic short-circuits with returns anyway, and immediately tried switching from your:if( cond ) { return x; } form to the statement modifier form: return x if cond;; and ... it made no difference. elsif still won out.

So then I looked at -MO=Deparse,-x9 to see if if/elsif/else chains were perhaps being converted to some kind of jump table or switch.

But no. They appear to be implemented as (the equivalent of) a chain of nested ternaries, but when I tried coding them that way, they were still slower than the if/elsif/else chain?

I then turned my attention to your rather more interesting innovation; namely (partially) unrolling a binary search. I tried adding another level, and got greater benefit; then another and (reduced, but) still more; so then another; and the benefits ran out :(

My (long and winding) benchmark and results are here:

Many thanks for a) responding; b) teaching me something(if/elsif); c) stimulating my brain in a different direction (hardcoding a binary search).


With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority". I'm with torvalds on this
In the absence of evidence, opinion is indistinguishable from prejudice. Agile (and TDD) debunked