Question reg Mysql Query Benchmark

I tried to respond to your query over email before, not sure if you received it.

So the plugin was updated and the new version uses different benchmarking algorithms. The old ones were having issues (some functions were removed/deprecated from PHP and mysql). So the numbers are explicitly NOT comparable between the old version and new.

The spirit is the same and 2.0 results will compare to 2.0 results in the graphs. 1.x results will compare against 1.x results (if you’re still running the old plugin). But 1.x and 2.0 results are apples-to-oranges different.

I am never thrilled to lose the history of benchmarks but when the functions used are no longer compatible with the latest versions of the software they are built on, it’s a breaking change that has to happen unfortunately.

Dear Kevin,

A big thank you for taking the time to reply so quickly to my message. I apologize somehow I didn’t get your email.

I completely understand what you’ve explained and I thank you for this great tools that you’ve developped for the WP Community.

Now, I’m clueless about why my Mysql Query Benchmark takes 4.686s which seems to me huge considering my stack and the efforts I’ve put into building it.

Would you be open to help me dig further ? My DevOps are trying to replicate your Mysql Query Benchmark directly but we are unable to understand the exact method.

Mysql Query takes nearly 4.8s for 1,000,000 ENCODE()s with a random seed.

We are on OLS stack (soon Litespeed Web Server) running PHP 8.1 + Redis object Cache.

Would you be open to work with us in order to find out if this value is accurate?

Thanks again Kevin.

Kind regards, Adrien

No problem, it’s probably helpful to post this all publicly for anyone else.

It is this line right here: https://github.com/kevinohashi/WPPerformanceTester/blob/426ffe4f91f8a39604ab8921e06cdf913b62a524/benchmark.php#L172

$query = “SELECT BENCHMARK(5000000, AES_ENCRYPT(CONCAT(‘WPHostingBenchmarks.com’,RAND()), UNHEX(SHA2(‘is part of Review Signal.com’,512))))”;

It’s running 5 million AES_ENCRYPT functions on your SQL server.

It is designed to take a while, and is 100% something that you would never do on a real production server. The goal is creating something heavy enough that we can measure the difference between systems. If everything handles something in 10ms, it’s harder to see a difference, and even harder as things go forward. I tried to create a function and run it enough times that it took a while. On purpose. That way it’s easier to see the difference between platforms and take into account things will likely get faster in the future.

I wouldn’t worry so much about the number itself (compare it against the benchmark of submitted scores) and compare it against different versions of your own setup. The actual raw number in isolation doesn’t mean a whole lot.

I hope that explains everything and helps you out. If not, please expand on your issue and I will try to answer.

Dear Kevin,

All understood. I really appreciate all your insights and clarifications.

Again a big thanks for your precious plugin and keeping it up to date.