mirror of
https://github.com/lovell/sharp.git
synced 2025-07-09 18:40:16 +02:00
d26f6b3b89
Add structure via mkdocs (replaces ever-growing README) Inline usage examples with the method they demonstrate Add changelog
2.2 KiB
2.2 KiB
Performance
Test environment
- AWS EC2 c4.xlarge
- Ubuntu 15.04
- Node.js 0.12.7
- libvips 8.0.2
- liborc 0.4.22
The contenders
- lwip 0.0.7 - Wrapper around CImg, compiles dependencies from source
- imagemagick-native git@45d4e2e - Wrapper around libmagick++, supports Buffers only.
- imagemagick 0.1.3 - Supports filesystem only and "has been unmaintained for a long time".
- gm 1.18.1 - Fully featured wrapper around GraphicsMagick's
convertcommand line utility. - sharp 0.11.0 - Caching within libvips disabled to ensure a fair comparison.
The task
Decompress a 2725x2225 JPEG image, resize to 720x480 using bilinear interpolation, then compress to JPEG.
Results
| Module | Input | Output | Ops/sec | Speed-up |
|---|---|---|---|---|
| lwip | file | file | 1.75 | 1 |
| lwip | buffer | buffer | 2.21 | 1.3 |
| imagemagick-native | buffer | buffer | 7.13 | 4.1 |
| gm | buffer | buffer | 7.27 | 4.2 |
| gm | file | file | 7.33 | 4.2 |
| imagemagick | file | file | 10.04 | 5.7 |
| sharp | stream | stream | 23.12 | 13.2 |
| sharp | file | file | 24.43 | 14.0 |
| sharp | file | buffer | 24.55 | 14.0 |
| sharp | buffer | file | 24.86 | 14.2 |
| sharp | buffer | buffer | 24.92 | 14.2 |
Greater performance can be expected with caching enabled (default) and using 8+ core machines.
The I/O limits of the relevant (de)compression library will generally determine maximum throughput.
Benchmark test prerequisites
Requires both ImageMagick and GraphicsMagick:
brew install imagemagick
brew install graphicsmagick
sudo apt-get install imagemagick graphicsmagick libmagick++-dev
Running the benchmark test
git clone https://github.com/lovell/sharp.git
cd sharp/test/bench
npm install
npm test