Hi,
I am on the PHP Project's Windows Team.
We build, optimize and extensively test the PHP builds for Windows (see: windows.php.net/downloads).
We now include Phalcon in our performance testing.
For the most recent 5.6, 7.0, 7.1 and 7.2 releases, see:
- https://windows.php.net/downloads/snaps/ostc/pftt/perf/PHP_7_2/7.2.5/results.html
- https://windows.php.net/downloads/snaps/ostc/pftt/perf/PHP_7_1/7.1.17/results.html
- https://windows.php.net/downloads/snaps/ostc/pftt/perf/PHP_7_0/7.0.30/results.html
- https://windows.php.net/downloads/snaps/ostc/pftt/perf/PHP_5_6/5.6.36/results.html
This is measured in Transactions Per Second(TPS); more is better. Servers are reset for each test run. Each test run has a warmup period(not counted) and then runs for 30 minutes at ~100% utilization. For Phalcon, I took the INVO app and modified it to list the companies in a table on the index page which our test then loads (so its a realistic test of MySQLi, ORM and Volt).
Phalcon far exceeds the performance of any application or framework we test (for internal benchmarks we do test PHP on Linux vs PHP on Windows - and all apps run with at least the same performance and reliability on Windows).
With PHP 7.2 and Phalcon 3.3.2, I'm usally getting 750-770 TPS on 4 CPU Cores running on Windows 2016/IIS 10 and MySQL 5.7.