I've noticed even really small programs (1192 bytes) take a lot longer to download when I program the chip (ATMega16) via my AVR ISP MKII (90 seconds) than when I download the same program via the bootloader (about 15 seconds.) It's almost as if the direct option is programming the full 16K program space. Or, perhaps it's using a very slow transfer speed. I'd like to speed this up, but I don't know where to look. Anyone have any ideas?
The AVRISP mkII has some speed / bit delay option that apparently comes from the factory with a random setting (e.g. really slow). You can adjust it with AVR Studio.
If you're a non-windows user, you can also set it with avrdude. Sorry, I don't recall how to do it at the moment, but I had to do the same with mine and successfully used avrdude on the mac to speed things up.
Sorry, I should have specified I'm running on a Mac, so I'm not sure how to use AVRStudio. If someone knows how to set this some other way, please let me know. Thanks.