libgcc is currently contains pure C functions with simple (i.e. non-optimal
implementations). Allowing individual targets to provide their own versions of
the functions could boost performance.
i.e. comparing the libgcc C implementation of __popcountdi2 to the assembly
version provided in the Opteron optimization manual, we find that the AMD
supplied version is half the instructions, branchless, uses no data tables (and
thus causes no cache misses), and runs in about 1/3 to 1/2 less time.
--
Summary: libgcc could use some target specific optimizations
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: other
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: nmiell at comcast dot net
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21812
