On 5/23/24 11:52 PM, Richard Biener wrote:
This worked out so I pushed the change. The gcc.dg/vect/pr97428.c
test is FAILing on RISC-V (it still gets 0 SLP), because of missed
load permutations. I hope the followup reorg for the load side will
fix this. It also FAILs gcc.target/riscv/rvv/autovec/struct/struct_vect-4.c
which does excessive assembly scanning on many functions - I'll leave
this for target maintainers to update - there's one or two functions
which we now expect to SLP.
Yea, folks got a bit carried away with the scan body capability.
Someone will have to follow-up behind you and clean this up a bit.
Thanks for checking it agains the CI system. While it's a bit on the
slow side, we are finding its helping catch real issues and keeping the
testsuite cleaner WRT FAILs.
jeff