Re: Re: Micro optimisations can pay off, and needn't be a maintenance problem

[dws]

You optimize one thing, attribute access, by 25%, but you get an overall speed up of a factor of 6. How's that possible?

The benchmark test cases include overhead for object creation. That might skew the test results a bit.

Getters and setters aren't benchmarked individually. Instead, the benchmark tests call getters and setters an equal number of times. But in the Z Buffer Demo getters seem to predominate (based on a quick read the code). Intuitively, getters are going to be faster than setters, perhaps enough so to explain a factor of 6 improvement.

[pilgrim]

Well, my analysis in this case should be taken with a salt lick, but breaking it out into the original (overall) benchmark followed by one each create/set and create/get shows that the "get" case is roughly 33% (up to 40%) faster than the "set" case. It also shows that the "set" gets (2%-3%) more of an improvement out of this optimization than the "get" does.

Breaking it out further, and accessing the previously-set $thing1 through $thing4 rather than creating a new object each iteration, I see that the individual "get" and "set" are actually about 37% faster.

So object creation does factor in, and getters are faster. It's still "only" a 37% improvement, rather than sixfold, but it's not a unilateral 37%.