Anyone been following this thread (which has become split unfortunately)?:
The chap is asking for people to supply results for processors other than the omap, and I thought we might be the perfect people (assuming most people reading the forum have a Zaurus or a 770 & a Zaurus) as we have a variety of processor types (and possibly optimisations).
I wonder how much of a difference the in-lining makes though. Anyone have any ideas?
-finline-functions would in-line things, but presumably functions like memcpy() would have to be available in source form to be inlined (making it more difficult for the test program, but not for patching glibc)?
I need to rebuild my toolchains, so I won't be able to do much for a day or so, but if anyone else has a working toolchain I'd be interested to see the results.