On Thu, Dec 11, 2025 at 4:21 PM Robin Dapp <[email protected]> wrote: > > >> diff --git > >> a/gcc/testsuite/gcc.target/riscv/rvv/autovec/gather-scatter/strided_store-2.c > >> > >> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/gather-scatter/strided_store-2.c > >> index 4acac8f7efd..3745de1fd76 100644 > >> --- > >> a/gcc/testsuite/gcc.target/riscv/rvv/autovec/gather-scatter/strided_store-2.c > >> +++ > >> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/gather-scatter/strided_store-2.c > >> @@ -40,6 +40,6 @@ > >> > >> TEST_ALL (TEST_LOOP) > >> > >> -/* { dg-final { scan-tree-dump-times " \.MASK_LEN_STRIDED_STORE " 44 > >> "optimized" } } */ > >> +/* { dg-final { scan-tree-dump-times " \.MASK_LEN_STRIDED_STORE " 44 > >> "optimized" { xfail { any-opts "-mrvv-max-lmul=dynamic" > >> "-mrvv-max-lmul=m8" } } } } */ > > > > Why did this fail??? I thought we added pattern should be able to > > optimize/vectorize more pattern rather than fail? > > We actually vectorize more. Or rather, we unroll more, which results in 49 > strided stores. That's only for LMUL8, though.
Then I would suggest either adding one more scan-tree-dump-times for -mrvv-max-lmul=m8 / -mrvv-max-lmul=dynamic Or adding comment to describe that > > > -- > Regards > Robin >
