https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125508
Bug ID: 125508
Summary: speed up cl_optimization_hash
Product: gcc
Version: 16.0
Status: UNCONFIRMED
Keywords: compile-time-hog
Severity: normal
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: pinskia at gcc dot gnu.org
Target Milestone: ---
Target: aarch64
I was running the GCC testsuite on aarch64-linux-gnu (this is still with
checking turned on) and looking into the top functions and
iterative_hash_host_wide_int was the top function since cl_optimization_hash
calls it.
I think we should be using add_int here rather than add_hwint for most of
these.
That should speed up things slightly because the upper 32bits will be 0.
In fact most of these are going to be 0 or 1 even. So maybe there is a way to
optimize this even more.
