Benchmark for Catberry's server-side rendering

Description

For now, it's just a comparison with bare express application which uses Jade template engine.

Methodology

The ApacheBench, Version 2.3 <$Revision: 1604373 $> is used for testing:

• Number of requests is 3000
• Concurrency level is 10

Both applications should:

• Render almost identical HTML (except Catberry renders the extra symbol in style tag).
• Use the same set of data generated by ./generate-data.js script and saved as ./data.json
• Have configurations with turned off log messages for every request
• Have last versions of packages (./setup.sh to install all dependencies)

Results

You can find the results in the ./results directory which are received on such configuration:

• MacBook Pro (Retina, 13-inch, Early 2015)
• 2,7 GHz Intel Core i5
• 8 GB 1867 MHz DDR3
• Intel Iris Graphics 6100 1536 MB
• OS X 10.11.3
• [email protected]
• [email protected]

node --version
v5.8.0

Here are the main numbers:

Catberry

Requests per second:    354.02 [#/sec] (mean)
Time per request:       28.247 [ms] (mean)
Time per request:       2.825 [ms] (mean, across all concurrent requests)
Transfer rate:          641.66 [Kbytes/sec] received

Express

Requests per second:    154.88 [#/sec] (mean)
Time per request:       64.566 [ms] (mean)
Time per request:       6.457 [ms] (mean, across all concurrent requests)
Transfer rate:          289.49 [Kbytes/sec] received

Conclusion

Despite on Catberry's stream-based and component-based rendering engine which has overhead related to usage of stream transformations with HTML parsing it has better performance than a bare Express with the same template engine at least on small size of data.

Contributing

Feel free to submit a PR with your benchmark built the same way.