optimizations involving IBM PC memory models

Can anyone tell me of any papers, books, or personal experiences of
optimization techniques relevant to the choice of memory models for IBM
PCs?


I have been toying with the idea that there are some optimizations
possible, such as using static analysis to detect when a pointer need only
be a near pointer, or when a function call need only be a near call, etc.
It seems to me that common compilers accept the user's choice of model and
don't attempt to determine whether it is an inefficient choice.


Thanks in advance,
David Spuler
[Some compilers automatically make calls to static functions near, even in
medium or large model code, but I've never seen one try to guess what pointer
model to use. -John]
--

Return to the comp.compilers page.
Search the comp.compilers archives again.