https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117558
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> --- Same reason for the aarch64 fails FAIL: gcc.target/aarch64/sve/struct_vect_22.c scan-assembler-times \\tld1d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_22.c scan-assembler-times \\tld4d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_22.c scan-assembler-times \\tst1d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_22.c scan-assembler-times \\tstr\\td 1 FAIL: gcc.target/aarch64/sve/struct_vect_23.c scan-assembler-times \\tld1d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_23.c scan-assembler-times \\tld4d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_23.c scan-assembler-times \\tst1d\\t 1 FAIL: gcc.target/aarch64/sve/struct_vect_23.c scan-assembler-times \\tb[.a-z]+\\t 12 and FAIL: gcc.target/aarch64/sve/strided_load_4.c scan-assembler-times \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+.s, sxtw 2\\]\\n 15 FAIL: gcc.target/aarch64/sve/strided_load_4.c scan-assembler-times \\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+.d, lsl 3\\]\\n 15 FAIL: gcc.target/aarch64/sve/strided_load_5.c scan-assembler-times \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+.s, uxtw\\]\\n 12 FAIL: gcc.target/aarch64/sve/strided_load_5.c scan-assembler-times \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+.s, sxtw\\]\\n 3 FAIL: gcc.target/aarch64/sve/strided_load_5.c scan-assembler-times \\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+.d\\]\\n 15 FAIL: gcc.target/aarch64/sve/strided_load_6.c scan-assembler-times \\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+.d\\]\\n 15 FAIL: gcc.target/aarch64/sve/strided_load_7.c scan-assembler-times \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+.s, sxtw\\]\\n 3 FAIL: gcc.target/aarch64/sve/strided_load_7.c scan-assembler-times \\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+.d\\]\\n 15 in theory we could - for single element - fall back to VMAT_ELEMENTWISE and then VMAT_GATHER_SCATTER for these cases.
