This is the mail archive of the
mailing list for the GCC project.
Re: [patch] cse.c: Speed up cse_reg_info maintanance - Part 3
> The fact remains, this is exactly the kinda of patch that kills
> performance on darwin.
> What we would like is imagine allocations that look like this:
> we want ln(N) calls to xmalloc, no reallocs, and not N calls to xmalloc.
I don't know much about darwin. If what you are saying is true, how
shall we handle many other passes that allocate and free pass-specific
memory? Should we really use this keep-and-grow idea everywhere in
GCC? Or should we wrap or replace xmalloc and free with something
that handles common allocation patterns?