|
|
Benchmarks
When Bundle-Fu was released, I made the statement that JS compression yields marginal results. Bundling yields about 6 times more performance than bundling+compression.
I still hold my position that bundling is the biggest speedup you can perform (knocked off 1 2/3 seconds load time in this benchmark). Javascript minimization only yielded an additional 1/7th of a second speedup. However, that's not to say it's a bad idea to compress javascript - every bit helps.
Here are some benchmarks of real world performance increases experienced by employing bundle-fu:
Overview
http://spinstudentlife.com/sightings
- Javascripts: 8 (192k) + 2 externals
- CSS: 4 (24k)
- Images: 11 (165k)
- HTML: (7k)
Test was performed by clearing the cache before each load. Load times calculated by firebug.
Benchmarks
Unbundled, Uncompressed
| Run | Site load time |
| 1 | 5.32 |
| 2 | 5.35 |
| 3 | 5.08 |
| 4 | 4.87 |
| 5 | 5.81 |
| Average | 5.29 |
Bundled, Uncompressed
JS size - 192k
| Run | JS load time | Site load time |
| 1 | 1.59 | 3.27 |
| 2 | 2.53 | 4.30 |
| 3 | 1.99 | 3.64 |
| 4 | 1.60 | 3.18 |
| 5 | 2.08 | 3.67 |
| Average | 1.96 | 3.61 |
Bundled, Compressed
JS size - 142k
| Run | JS load time | Site load time |
| 1 | 1.78 | 3.28 |
| 2 | 2.03 | 3.63 |
| 3 | 1.35 | 3.42 |
| 4 | 1.51 | 3.17 |
| 5 | 1.89 | 3.81 |
| Average | 1.71 | 3.46 |
Summary
Overall speed gained by:
| bundling | 3.61s vs 5.29s | 46% |
| bundling and compressing | 3.46s vs 5.29s | 53% |
Further speed could be obtained by implementing server-side gzip compression of css/js assets.
Sign in to add a comment
Using bundle_fu didn't improve performance at my side (apache bench). 100 Request, 10 Concurrent, single mongrel instance, 5 JS + 3 CSS
Is there something I'm missing?
bundle-fu optimizes overall loading time, not page rendering speed. Use firebug to measure your speedup.