Most (all?) other compilers generally have canned start-up code which include all routines to clear & initialize data, configure the processor, etc... which can be quite large. A new technology we developed for our compilers (included in our Holtek compiler) allows the compiler to write custom start-up code. Our compiler will analyze the program being compiled and "write" custom code for exactly what that program needs. Since it knows exactly how much data needs clearing/initializing it will use an optimal sequence of instructions to do so. For example, for clearing (zeroing) data, if there are is a lot to do it will probably create a loop. On the other hand, if there are only a few bytes, it may simply put down a sequence of clear instructions. In a minimal case there may be no start-up code at all and the main function is simply linked at the reset vector.