Filling Large Arrays

Anon

Is there a faster way to create these arrays on the fly?

srand ((double) microtime() * 1000000);
for ($ii = 1; $ii <= 1000; $ii++) {
array_push($words, #[ 1000 - 12 random character elements ]
chr(rand(1,26)+65) . chr(rand(1,26)+65) . chr(rand(1,26)+65) .
chr(rand(1,26)+65) . chr(rand(1,26)+65) . chr(rand(1,26)+65) .
chr(rand(1,26)+65) . chr(rand(1,26)+65) . chr(rand(1,26)+65) .
chr(rand(1,26)+65) . chr(rand(1,26)+65) . chr(rand(1,26)+65));
array_push($nums, #[ 1000 - 12 random integer elements ]
rand(0,26) . rand(0,26) . rand(0,26) .
rand(0,26) . rand(0,26) . rand(0,26));
}

Anon

We will start with your "for" loop:

for ($ii = 1; $ii <= 1000; $ii++) {

becomes:

for($ii = 1000; --$ii > 0😉

that will make your "for" loop faster, as comparing to zero is much faster than "<= 1000".

2nd:
You here use array_push to add the element to the end of the array:

array_push($words, chr(rand(1,26)+65).....);

I'm not sure if this is true, but I would think that 1000 function calls must be slower than 1000:

$words[] = chr(rand(1,26)+65).....;

This does the same thing without the 1000 function calls.

3rd:
chr(rand(1,26)+65)

I'm sure adding 65 to a number doesn't take long, but why not use chr(rand(66,91))? The 2 params to rand are a min and a max so the outcome will be identical.

In programming terms "rand(1, 26)+65" == "rand(66, 91)" 😉

Hope this helps,
Jack

kparker

as comparing to zero is much faster than "<= 1000".

Have you ever measured the difference with PHP?

Anon

Thanks Jack. Those tips are duly noted.

I am actually working on a benchmarking suite to compare web development languages and I have been humming and hawing over how to fairly and accurately represent each language. I decided that the best way would be to write the fastest possible code for each language, but this doesn't quite work in some situations.

Thanks muchly for your tips.