I’ve become increasingly aware of a big difference in the uncompressed image sizes between debug and release - for my ME9210 project its now the difference between 1570K and 2360K - or nearly 800K.
I can understand that in debug mode extra code is needed to initialise the I/O, for example - but that doesn’t explain the very large difference.
Any thoughts?
are you comapring image.bin to image.bin? Perhaps otimization is off for debug one?
I’m comparing the two uncompressed image sizes (i.e. generated code sizes).
Optimisation is set to the Digi defaults for debug and release.
I’ve just compiled the release image with optimisation off (to match that used during debug), and get an uncompressed image size of 1832K - so a little bit bigger than the optimised release image, but still way smaller than the debug image.
Using latest Net+O/S 7.4.2.5, incidentally
What files exactl;y are you comparing? image.elf? you should really be comapring image.bin to image.bin, this is raw code with now debug information.
On completion of a make, the console window gives two file sizes - the uncompressed file size (i.e. the ‘raw’ code image size) and the compressed (image.bin) size. I’m quoting the first of these numbers.