µMonitor strikes back

Given the recent interest on my torrent-monitoring application (thank you to the guys at torrentfreak.com for finding it out and writing about it), I announce that I will start development of the application once more to try and satisfy user requests (and fix a few bugs that I know are around).

Stay tuned for an update, which I hope it will arrive as soon as possible (at the current time I’ve been focusing more on other applications!)

Thanks for the interest, and remember that if you enjoy the application you might donate, it will keep me updating (and not look into other forms of financing, such as in-ads, which I hate) it with more than just self-interest for it!

ciop ciop

Benchmarks running on the simulator and on the device

ImageMagick on iPhone – Benchmarks

Thanks to Karl (see previous post comments and update) the XCode project has now the possibility to work on the images uncompressed. We (me and Karl) wanted to see the difference for ImageMagick to work with a compressed format (JPEG, for instance) and with an uncompressed format (raw data).

I’ve added to the project a simple benchmark, consisting in running the MagickWand creation, filtering and destruction a given number of times (customizable in the beginning of the source file, for instance 10), calculating how long it takes each time and in total.

The results are impressive, and can be better exaplined by looking at this simple graph:

Benchmarks running on the simulator and on the device

Benchmarks running on the simulator and on the device (click to see better!)

As you can see working with uncompressed data achieves 3x faster results on the iPhone device, with a mean running time of 0.85 seconds to run an ordered posterize filter on an image of size 320×460 (the size of the iPhone screen). Similar results are also on the Simulator achieving 3.7x faster filtering for the same image.

The end line is trivial, working with uncompressed data, while being less easy (but as you can see from the project code not extreme) or intuitive yields much faster results!

You can download a better looking graph and the IM_test project as usual.

For any comment don’t hesitate to write, as you’ve seen I try to pick up my comments as much as I can (even on holiday!)

Thanks goes again to Karl who’s has the idea of working with raw images, and provided with the code to achieve so.

ciop ciop