PHP 7 word count benchmark (opens in new tab)

(juditacs.github.io)

13 pointsclemlais10y ago5 comments

5 comments

The algorithms differ considerably across language and the PHP one is particularly inefficient. The algorithms must be held constant in order for this "benchmark" to be valid.

As of right now this is just a programming competition across a variety of languages.

klodolph10y ago

I'm imagining an alternate version of the Wizard of Oz, where Dorothy clicks her ruby keyboard and says, "All benchmarks are flawed. All benchmarks are flawed. All benchmarks are flawed."

This is not an uncommon technique for comparing programming languages. We are not just trying to compare the performance of different implementations, but we are trying to compare the performance of different idiomatic programming styles in different languages. The shootout does the same thing. And yes, the shootout is also flawed.

NEDM6410y ago

Surprised by Python and Go results. They should have tested with PyPy too.

rurban10y ago

It's only testing the quality of hash tables and sort, not much else. siphash beating FNV1 can only be explained with the table implementation

leeoniya10y ago

i wonder if you could do even better with PHP's SPL* functions like:

http://php.net/manual/en/class.splfixedarray.php

j / k navigate · click thread line to collapse