Tool/software:
In our product, we support over-the-air updates to the firmware contained in our MSP430 microprocessors.
Today, we ship each new firmware image as single, contiguous binary image. This is pretty good, getting us an ~3 times advantage compared to shipping the original TI .txt files produced by the IAR XLinker.
But we'd like to do better. We experimented with LZMA and while it definitely did a good job of compressing the images, the decompressor itself represents a lot of code and executing on the MSP430, the LZMA decompressor was pretty slow.
So we're wondering if there's a better trade-off between compression ration, size of the decompressor, and execution speed.
I've looked at the statistics of some typical MSP430 images and it looks like they'd be somewhat susceptible to some simple compression algorithms but before I dive in and write code, I'm wondering if anyone else has already trodden across this ground.
Does anyone have any thoughts regarding simple compression algorithms for MSP430 program images?
