Hi, I'd like to ping this patch as well as its sblings. Thanks in advance.
1/4 v3 https://gcc.gnu.org/pipermail/gcc-patches/2020-February/540171.html 2/4 v3 https://gcc.gnu.org/pipermail/gcc-patches/2020-March/541387.html 3/4 v3 https://gcc.gnu.org/pipermail/gcc-patches/2020-May/545643.html BR, Kewen on 2020/3/3 下午8:25, Kewen.Lin wrote: > Hi Richard, > > Thanks for your comments! It's a good idea to use param due to the > flexibility. And yes, it sounds good to have more targets to try and > make it better. But I have a bit concern on turning it on by default. > Since it replies on unroll factor estimation, as part 1/4 shows, it > calls targetm.loop_unroll_adjust if target supports, which used to > work on RTL level. To avoid possible ICE, I'm intended to turn it > off for those targets (s390 & i386) with that hook, since without good > understanding on those targets, it's hard for me to extend them with > gimple level support. Does it make sense? > > The updated patch has been attached. > > BR, > Kewen > --------- > > gcc/ChangeLog > > 2020-03-03 Kewen Lin <li...@gcc.gnu.org> > > * doc/invoke.texi (iv-consider-reg-offset-for-unroll): Document new > option. > * params.opt (iv-consider-reg-offset-for-unroll): New. > * config/s390/s390.c (s390_option_override_internal): Disable parameter > iv-consider-reg-offset-for-unroll by default. > * config/i386/i386-options.c (ix86_option_override_internal): Likewise. >
