Home | History | Annotate | Download | only in X86

Lines Matching full:shuffle

11 ; target-specific shuffle instructions where reasonable.
273 ; Verify that DAGCombiner moves the shuffle after the xor/and/or even if shuffles
985 ; Verify that we don't optimize the following cases. We expect more than one shuffle.
988 ; making sense with th enew vector shuffle lowering. Revisit at least testing for
1227 ; shuffle(shuffle(x, y), undef) -> shuffle(y, undef)
2132 %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 undef, i32 2, i32 3>
2133 %shuffle6 = shufflevector <4 x float> %shuffle, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 3>
2161 %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 1, i32 undef, i32 3>
2162 %shuffle6 = shufflevector <4 x float> %shuffle, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
2188 %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 undef, i32 undef>
2189 %shuffle6 = shufflevector <4 x float> %shuffle, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 undef>
2244 ; (shuffle(shuffle A, Undef, M0), B, M1) -> (shuffle A, B, M2)
2346 ; (shuffle(shuffle A, Undef, M0), A, M1) -> (shuffle A, Undef, M2)
2531 ; shuffle(B, shuffle(A, Undef)) -> shuffle(shuffle(A, Undef), B)
2534 ; (shuffle(shuffle A, Undef, M0), A, M1) -> (shuffle A, Undef, M2)
2536 ; As a result, all the shuffle pairs in each function below should be
2537 ; combined into a single legal shuffle operation.
2623 ; operations feeding into a shuffle. The AVX cases are the important ones as