mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-25 20:23:11 +01:00
2e307633f3
Re-recommiting after landing DAG extension-crash fix. Recommiting after adding check to avoid miscomputing alias information on addresses of the same base but different subindices. Memory accesses offset from frame indices may alias, e.g., we may merge write from function arguments passed on the stack when they are contiguous. As a result, when checking aliasing, we consider the underlying frame index's offset from the stack pointer. Static allocs are realized as stack objects in SelectionDAG, but its offset is not set until post-DAG causing DAGCombiner's alias check to consider access to static allocas to frequently alias. Modify isAlias to consider access between static allocas and access from other frame objects to be considered aliasing. Many test changes are included here. Most are fixes for tests which indirectly relied on our aliasing ability and needed to be modified to preserve their original intent. The remaining tests have minor improvements due to relaxed ordering. The exception is CodeGen/X86/2011-10-19-widen_vselect.ll which has a minor degradation dispite though the pre-legalized DAG is improved. Reviewers: rnk, mkuper, jonpa, hfinkel, uweigand Reviewed By: rnk Subscribers: sdardis, nemanjai, javed.absar, llvm-commits Differential Revision: https://reviews.llvm.org/D33345 llvm-svn: 308350
465 lines
19 KiB
LLVM
465 lines
19 KiB
LLVM
; RUN: llc < %s -mtriple=arm64-eabi -aarch64-neon-syntax=apple | FileCheck %s
|
|
|
|
define void @test_vext_s8() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_s8:
|
|
; CHECK: {{ext.8.*#1}}
|
|
%xS8x8 = alloca <8 x i8>, align 8
|
|
%__a = alloca <8 x i8>, align 8
|
|
%__b = alloca <8 x i8>, align 8
|
|
%tmp = load <8 x i8>, <8 x i8>* %xS8x8, align 8
|
|
store <8 x i8> %tmp, <8 x i8>* %__a, align 8
|
|
%tmp1 = load <8 x i8>, <8 x i8>* %xS8x8, align 8
|
|
store <8 x i8> %tmp1, <8 x i8>* %__b, align 8
|
|
%tmp2 = load <8 x i8>, <8 x i8>* %__a, align 8
|
|
%tmp3 = load <8 x i8>, <8 x i8>* %__b, align 8
|
|
%vext = shufflevector <8 x i8> %tmp2, <8 x i8> %tmp3, <8 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
|
|
store <8 x i8> %vext, <8 x i8>* %xS8x8, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_u8() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_u8:
|
|
; CHECK: {{ext.8.*#2}}
|
|
%xU8x8 = alloca <8 x i8>, align 8
|
|
%__a = alloca <8 x i8>, align 8
|
|
%__b = alloca <8 x i8>, align 8
|
|
%tmp = load <8 x i8>, <8 x i8>* %xU8x8, align 8
|
|
store <8 x i8> %tmp, <8 x i8>* %__a, align 8
|
|
%tmp1 = load <8 x i8>, <8 x i8>* %xU8x8, align 8
|
|
store <8 x i8> %tmp1, <8 x i8>* %__b, align 8
|
|
%tmp2 = load <8 x i8>, <8 x i8>* %__a, align 8
|
|
%tmp3 = load <8 x i8>, <8 x i8>* %__b, align 8
|
|
%vext = shufflevector <8 x i8> %tmp2, <8 x i8> %tmp3, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
|
|
store <8 x i8> %vext, <8 x i8>* %xU8x8, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_p8() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_p8:
|
|
; CHECK: {{ext.8.*#3}}
|
|
%xP8x8 = alloca <8 x i8>, align 8
|
|
%__a = alloca <8 x i8>, align 8
|
|
%__b = alloca <8 x i8>, align 8
|
|
%tmp = load <8 x i8>, <8 x i8>* %xP8x8, align 8
|
|
store <8 x i8> %tmp, <8 x i8>* %__a, align 8
|
|
%tmp1 = load <8 x i8>, <8 x i8>* %xP8x8, align 8
|
|
store <8 x i8> %tmp1, <8 x i8>* %__b, align 8
|
|
%tmp2 = load <8 x i8>, <8 x i8>* %__a, align 8
|
|
%tmp3 = load <8 x i8>, <8 x i8>* %__b, align 8
|
|
%vext = shufflevector <8 x i8> %tmp2, <8 x i8> %tmp3, <8 x i32> <i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10>
|
|
store <8 x i8> %vext, <8 x i8>* %xP8x8, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_s16() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_s16:
|
|
; CHECK: {{ext.8.*#2}}
|
|
%xS16x4 = alloca <4 x i16>, align 8
|
|
%__a = alloca <4 x i16>, align 8
|
|
%__b = alloca <4 x i16>, align 8
|
|
%tmp = load <4 x i16>, <4 x i16>* %xS16x4, align 8
|
|
store <4 x i16> %tmp, <4 x i16>* %__a, align 8
|
|
%tmp1 = load <4 x i16>, <4 x i16>* %xS16x4, align 8
|
|
store <4 x i16> %tmp1, <4 x i16>* %__b, align 8
|
|
%tmp2 = load <4 x i16>, <4 x i16>* %__a, align 8
|
|
%tmp3 = bitcast <4 x i16> %tmp2 to <8 x i8>
|
|
%tmp4 = load <4 x i16>, <4 x i16>* %__b, align 8
|
|
%tmp5 = bitcast <4 x i16> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <4 x i16>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <4 x i16>
|
|
%vext = shufflevector <4 x i16> %tmp6, <4 x i16> %tmp7, <4 x i32> <i32 1, i32 2, i32 3, i32 4>
|
|
store <4 x i16> %vext, <4 x i16>* %xS16x4, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_u16() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_u16:
|
|
; CHECK: {{ext.8.*#4}}
|
|
%xU16x4 = alloca <4 x i16>, align 8
|
|
%__a = alloca <4 x i16>, align 8
|
|
%__b = alloca <4 x i16>, align 8
|
|
%tmp = load <4 x i16>, <4 x i16>* %xU16x4, align 8
|
|
store <4 x i16> %tmp, <4 x i16>* %__a, align 8
|
|
%tmp1 = load <4 x i16>, <4 x i16>* %xU16x4, align 8
|
|
store <4 x i16> %tmp1, <4 x i16>* %__b, align 8
|
|
%tmp2 = load <4 x i16>, <4 x i16>* %__a, align 8
|
|
%tmp3 = bitcast <4 x i16> %tmp2 to <8 x i8>
|
|
%tmp4 = load <4 x i16>, <4 x i16>* %__b, align 8
|
|
%tmp5 = bitcast <4 x i16> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <4 x i16>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <4 x i16>
|
|
%vext = shufflevector <4 x i16> %tmp6, <4 x i16> %tmp7, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
|
|
store <4 x i16> %vext, <4 x i16>* %xU16x4, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_p16() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_p16:
|
|
; CHECK: {{ext.8.*#6}}
|
|
%xP16x4 = alloca <4 x i16>, align 8
|
|
%__a = alloca <4 x i16>, align 8
|
|
%__b = alloca <4 x i16>, align 8
|
|
%tmp = load <4 x i16>, <4 x i16>* %xP16x4, align 8
|
|
store <4 x i16> %tmp, <4 x i16>* %__a, align 8
|
|
%tmp1 = load <4 x i16>, <4 x i16>* %xP16x4, align 8
|
|
store <4 x i16> %tmp1, <4 x i16>* %__b, align 8
|
|
%tmp2 = load <4 x i16>, <4 x i16>* %__a, align 8
|
|
%tmp3 = bitcast <4 x i16> %tmp2 to <8 x i8>
|
|
%tmp4 = load <4 x i16>, <4 x i16>* %__b, align 8
|
|
%tmp5 = bitcast <4 x i16> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <4 x i16>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <4 x i16>
|
|
%vext = shufflevector <4 x i16> %tmp6, <4 x i16> %tmp7, <4 x i32> <i32 3, i32 4, i32 5, i32 6>
|
|
store <4 x i16> %vext, <4 x i16>* %xP16x4, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_s32() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_s32:
|
|
; CHECK: {{rev64.2s.*}}
|
|
%xS32x2 = alloca <2 x i32>, align 8
|
|
%__a = alloca <2 x i32>, align 8
|
|
%__b = alloca <2 x i32>, align 8
|
|
%tmp = load <2 x i32>, <2 x i32>* %xS32x2, align 8
|
|
store <2 x i32> %tmp, <2 x i32>* %__a, align 8
|
|
%tmp1 = load <2 x i32>, <2 x i32>* %xS32x2, align 8
|
|
store <2 x i32> %tmp1, <2 x i32>* %__b, align 8
|
|
%tmp2 = load <2 x i32>, <2 x i32>* %__a, align 8
|
|
%tmp3 = bitcast <2 x i32> %tmp2 to <8 x i8>
|
|
%tmp4 = load <2 x i32>, <2 x i32>* %__b, align 8
|
|
%tmp5 = bitcast <2 x i32> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <2 x i32>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <2 x i32>
|
|
%vext = shufflevector <2 x i32> %tmp6, <2 x i32> %tmp7, <2 x i32> <i32 1, i32 2>
|
|
store <2 x i32> %vext, <2 x i32>* %xS32x2, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_u32() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_u32:
|
|
; CHECK: {{rev64.2s.*}}
|
|
%xU32x2 = alloca <2 x i32>, align 8
|
|
%__a = alloca <2 x i32>, align 8
|
|
%__b = alloca <2 x i32>, align 8
|
|
%tmp = load <2 x i32>, <2 x i32>* %xU32x2, align 8
|
|
store <2 x i32> %tmp, <2 x i32>* %__a, align 8
|
|
%tmp1 = load <2 x i32>, <2 x i32>* %xU32x2, align 8
|
|
store <2 x i32> %tmp1, <2 x i32>* %__b, align 8
|
|
%tmp2 = load <2 x i32>, <2 x i32>* %__a, align 8
|
|
%tmp3 = bitcast <2 x i32> %tmp2 to <8 x i8>
|
|
%tmp4 = load <2 x i32>, <2 x i32>* %__b, align 8
|
|
%tmp5 = bitcast <2 x i32> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <2 x i32>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <2 x i32>
|
|
%vext = shufflevector <2 x i32> %tmp6, <2 x i32> %tmp7, <2 x i32> <i32 1, i32 2>
|
|
store <2 x i32> %vext, <2 x i32>* %xU32x2, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_f32() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_f32:
|
|
; CHECK: {{rev64.2s.*}}
|
|
%xF32x2 = alloca <2 x float>, align 8
|
|
%__a = alloca <2 x float>, align 8
|
|
%__b = alloca <2 x float>, align 8
|
|
%tmp = load <2 x float>, <2 x float>* %xF32x2, align 8
|
|
store <2 x float> %tmp, <2 x float>* %__a, align 8
|
|
%tmp1 = load <2 x float>, <2 x float>* %xF32x2, align 8
|
|
store <2 x float> %tmp1, <2 x float>* %__b, align 8
|
|
%tmp2 = load <2 x float>, <2 x float>* %__a, align 8
|
|
%tmp3 = bitcast <2 x float> %tmp2 to <8 x i8>
|
|
%tmp4 = load <2 x float>, <2 x float>* %__b, align 8
|
|
%tmp5 = bitcast <2 x float> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <2 x float>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <2 x float>
|
|
%vext = shufflevector <2 x float> %tmp6, <2 x float> %tmp7, <2 x i32> <i32 1, i32 2>
|
|
store <2 x float> %vext, <2 x float>* %xF32x2, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_s64() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_s64:
|
|
; CHECK_FIXME: {{rev64.2s.*}}
|
|
; this just turns into a load of the second element
|
|
%xS64x1 = alloca <1 x i64>, align 8
|
|
%__a = alloca <1 x i64>, align 8
|
|
%__b = alloca <1 x i64>, align 8
|
|
%tmp = load <1 x i64>, <1 x i64>* %xS64x1, align 8
|
|
store <1 x i64> %tmp, <1 x i64>* %__a, align 8
|
|
%tmp1 = load <1 x i64>, <1 x i64>* %xS64x1, align 8
|
|
store <1 x i64> %tmp1, <1 x i64>* %__b, align 8
|
|
%tmp2 = load <1 x i64>, <1 x i64>* %__a, align 8
|
|
%tmp3 = bitcast <1 x i64> %tmp2 to <8 x i8>
|
|
%tmp4 = load <1 x i64>, <1 x i64>* %__b, align 8
|
|
%tmp5 = bitcast <1 x i64> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <1 x i64>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <1 x i64>
|
|
%vext = shufflevector <1 x i64> %tmp6, <1 x i64> %tmp7, <1 x i32> <i32 1>
|
|
store <1 x i64> %vext, <1 x i64>* %xS64x1, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vext_u64() nounwind ssp {
|
|
; CHECK-LABEL: test_vext_u64:
|
|
; CHECK_FIXME: {{ext.8.*#1}}
|
|
; this is turned into a simple load of the 2nd element
|
|
%xU64x1 = alloca <1 x i64>, align 8
|
|
%__a = alloca <1 x i64>, align 8
|
|
%__b = alloca <1 x i64>, align 8
|
|
%tmp = load <1 x i64>, <1 x i64>* %xU64x1, align 8
|
|
store <1 x i64> %tmp, <1 x i64>* %__a, align 8
|
|
%tmp1 = load <1 x i64>, <1 x i64>* %xU64x1, align 8
|
|
store <1 x i64> %tmp1, <1 x i64>* %__b, align 8
|
|
%tmp2 = load <1 x i64>, <1 x i64>* %__a, align 8
|
|
%tmp3 = bitcast <1 x i64> %tmp2 to <8 x i8>
|
|
%tmp4 = load <1 x i64>, <1 x i64>* %__b, align 8
|
|
%tmp5 = bitcast <1 x i64> %tmp4 to <8 x i8>
|
|
%tmp6 = bitcast <8 x i8> %tmp3 to <1 x i64>
|
|
%tmp7 = bitcast <8 x i8> %tmp5 to <1 x i64>
|
|
%vext = shufflevector <1 x i64> %tmp6, <1 x i64> %tmp7, <1 x i32> <i32 1>
|
|
store <1 x i64> %vext, <1 x i64>* %xU64x1, align 8
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_s8() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_s8:
|
|
; CHECK: {{ext.16.*#4}}
|
|
%xS8x16 = alloca <16 x i8>, align 16
|
|
%__a = alloca <16 x i8>, align 16
|
|
%__b = alloca <16 x i8>, align 16
|
|
%tmp = load <16 x i8>, <16 x i8>* %xS8x16, align 16
|
|
store <16 x i8> %tmp, <16 x i8>* %__a, align 16
|
|
%tmp1 = load <16 x i8>, <16 x i8>* %xS8x16, align 16
|
|
store <16 x i8> %tmp1, <16 x i8>* %__b, align 16
|
|
%tmp2 = load <16 x i8>, <16 x i8>* %__a, align 16
|
|
%tmp3 = load <16 x i8>, <16 x i8>* %__b, align 16
|
|
%vext = shufflevector <16 x i8> %tmp2, <16 x i8> %tmp3, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19>
|
|
store <16 x i8> %vext, <16 x i8>* %xS8x16, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_u8() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_u8:
|
|
; CHECK: {{ext.16.*#5}}
|
|
%xU8x16 = alloca <16 x i8>, align 16
|
|
%__a = alloca <16 x i8>, align 16
|
|
%__b = alloca <16 x i8>, align 16
|
|
%tmp = load <16 x i8>, <16 x i8>* %xU8x16, align 16
|
|
store <16 x i8> %tmp, <16 x i8>* %__a, align 16
|
|
%tmp1 = load <16 x i8>, <16 x i8>* %xU8x16, align 16
|
|
store <16 x i8> %tmp1, <16 x i8>* %__b, align 16
|
|
%tmp2 = load <16 x i8>, <16 x i8>* %__a, align 16
|
|
%tmp3 = load <16 x i8>, <16 x i8>* %__b, align 16
|
|
%vext = shufflevector <16 x i8> %tmp2, <16 x i8> %tmp3, <16 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20>
|
|
store <16 x i8> %vext, <16 x i8>* %xU8x16, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_p8() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_p8:
|
|
; CHECK: {{ext.16.*#6}}
|
|
%xP8x16 = alloca <16 x i8>, align 16
|
|
%__a = alloca <16 x i8>, align 16
|
|
%__b = alloca <16 x i8>, align 16
|
|
%tmp = load <16 x i8>, <16 x i8>* %xP8x16, align 16
|
|
store <16 x i8> %tmp, <16 x i8>* %__a, align 16
|
|
%tmp1 = load <16 x i8>, <16 x i8>* %xP8x16, align 16
|
|
store <16 x i8> %tmp1, <16 x i8>* %__b, align 16
|
|
%tmp2 = load <16 x i8>, <16 x i8>* %__a, align 16
|
|
%tmp3 = load <16 x i8>, <16 x i8>* %__b, align 16
|
|
%vext = shufflevector <16 x i8> %tmp2, <16 x i8> %tmp3, <16 x i32> <i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21>
|
|
store <16 x i8> %vext, <16 x i8>* %xP8x16, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_s16() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_s16:
|
|
; CHECK: {{ext.16.*#14}}
|
|
%xS16x8 = alloca <8 x i16>, align 16
|
|
%__a = alloca <8 x i16>, align 16
|
|
%__b = alloca <8 x i16>, align 16
|
|
%tmp = load <8 x i16>, <8 x i16>* %xS16x8, align 16
|
|
store <8 x i16> %tmp, <8 x i16>* %__a, align 16
|
|
%tmp1 = load <8 x i16>, <8 x i16>* %xS16x8, align 16
|
|
store <8 x i16> %tmp1, <8 x i16>* %__b, align 16
|
|
%tmp2 = load <8 x i16>, <8 x i16>* %__a, align 16
|
|
%tmp3 = bitcast <8 x i16> %tmp2 to <16 x i8>
|
|
%tmp4 = load <8 x i16>, <8 x i16>* %__b, align 16
|
|
%tmp5 = bitcast <8 x i16> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <8 x i16>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <8 x i16>
|
|
%vext = shufflevector <8 x i16> %tmp6, <8 x i16> %tmp7, <8 x i32> <i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14>
|
|
store <8 x i16> %vext, <8 x i16>* %xS16x8, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_u16() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_u16:
|
|
; CHECK: {{ext.16.*#8}}
|
|
%xU16x8 = alloca <8 x i16>, align 16
|
|
%__a = alloca <8 x i16>, align 16
|
|
%__b = alloca <8 x i16>, align 16
|
|
%tmp = load <8 x i16>, <8 x i16>* %xU16x8, align 16
|
|
store <8 x i16> %tmp, <8 x i16>* %__a, align 16
|
|
%tmp1 = load <8 x i16>, <8 x i16>* %xU16x8, align 16
|
|
store <8 x i16> %tmp1, <8 x i16>* %__b, align 16
|
|
%tmp2 = load <8 x i16>, <8 x i16>* %__a, align 16
|
|
%tmp3 = bitcast <8 x i16> %tmp2 to <16 x i8>
|
|
%tmp4 = load <8 x i16>, <8 x i16>* %__b, align 16
|
|
%tmp5 = bitcast <8 x i16> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <8 x i16>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <8 x i16>
|
|
%vext = shufflevector <8 x i16> %tmp6, <8 x i16> %tmp7, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
|
|
store <8 x i16> %vext, <8 x i16>* %xU16x8, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_p16() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_p16:
|
|
; CHECK: {{ext.16.*#10}}
|
|
%xP16x8 = alloca <8 x i16>, align 16
|
|
%__a = alloca <8 x i16>, align 16
|
|
%__b = alloca <8 x i16>, align 16
|
|
%tmp = load <8 x i16>, <8 x i16>* %xP16x8, align 16
|
|
store <8 x i16> %tmp, <8 x i16>* %__a, align 16
|
|
%tmp1 = load <8 x i16>, <8 x i16>* %xP16x8, align 16
|
|
store <8 x i16> %tmp1, <8 x i16>* %__b, align 16
|
|
%tmp2 = load <8 x i16>, <8 x i16>* %__a, align 16
|
|
%tmp3 = bitcast <8 x i16> %tmp2 to <16 x i8>
|
|
%tmp4 = load <8 x i16>, <8 x i16>* %__b, align 16
|
|
%tmp5 = bitcast <8 x i16> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <8 x i16>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <8 x i16>
|
|
%vext = shufflevector <8 x i16> %tmp6, <8 x i16> %tmp7, <8 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12>
|
|
store <8 x i16> %vext, <8 x i16>* %xP16x8, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_s32() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_s32:
|
|
; CHECK: {{ext.16.*#4}}
|
|
%xS32x4 = alloca <4 x i32>, align 16
|
|
%__a = alloca <4 x i32>, align 16
|
|
%__b = alloca <4 x i32>, align 16
|
|
%tmp = load <4 x i32>, <4 x i32>* %xS32x4, align 16
|
|
store <4 x i32> %tmp, <4 x i32>* %__a, align 16
|
|
%tmp1 = load <4 x i32>, <4 x i32>* %xS32x4, align 16
|
|
store <4 x i32> %tmp1, <4 x i32>* %__b, align 16
|
|
%tmp2 = load <4 x i32>, <4 x i32>* %__a, align 16
|
|
%tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
|
|
%tmp4 = load <4 x i32>, <4 x i32>* %__b, align 16
|
|
%tmp5 = bitcast <4 x i32> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <4 x i32>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <4 x i32>
|
|
%vext = shufflevector <4 x i32> %tmp6, <4 x i32> %tmp7, <4 x i32> <i32 1, i32 2, i32 3, i32 4>
|
|
store <4 x i32> %vext, <4 x i32>* %xS32x4, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_u32() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_u32:
|
|
; CHECK: {{ext.16.*#8}}
|
|
%xU32x4 = alloca <4 x i32>, align 16
|
|
%__a = alloca <4 x i32>, align 16
|
|
%__b = alloca <4 x i32>, align 16
|
|
%tmp = load <4 x i32>, <4 x i32>* %xU32x4, align 16
|
|
store <4 x i32> %tmp, <4 x i32>* %__a, align 16
|
|
%tmp1 = load <4 x i32>, <4 x i32>* %xU32x4, align 16
|
|
store <4 x i32> %tmp1, <4 x i32>* %__b, align 16
|
|
%tmp2 = load <4 x i32>, <4 x i32>* %__a, align 16
|
|
%tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
|
|
%tmp4 = load <4 x i32>, <4 x i32>* %__b, align 16
|
|
%tmp5 = bitcast <4 x i32> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <4 x i32>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <4 x i32>
|
|
%vext = shufflevector <4 x i32> %tmp6, <4 x i32> %tmp7, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
|
|
store <4 x i32> %vext, <4 x i32>* %xU32x4, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_f32() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_f32:
|
|
; CHECK: {{ext.16.*#12}}
|
|
%xF32x4 = alloca <4 x float>, align 16
|
|
%__a = alloca <4 x float>, align 16
|
|
%__b = alloca <4 x float>, align 16
|
|
%tmp = load <4 x float>, <4 x float>* %xF32x4, align 16
|
|
store <4 x float> %tmp, <4 x float>* %__a, align 16
|
|
%tmp1 = load <4 x float>, <4 x float>* %xF32x4, align 16
|
|
store <4 x float> %tmp1, <4 x float>* %__b, align 16
|
|
%tmp2 = load <4 x float>, <4 x float>* %__a, align 16
|
|
%tmp3 = bitcast <4 x float> %tmp2 to <16 x i8>
|
|
%tmp4 = load <4 x float>, <4 x float>* %__b, align 16
|
|
%tmp5 = bitcast <4 x float> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <4 x float>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <4 x float>
|
|
%vext = shufflevector <4 x float> %tmp6, <4 x float> %tmp7, <4 x i32> <i32 3, i32 4, i32 5, i32 6>
|
|
store <4 x float> %vext, <4 x float>* %xF32x4, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_s64() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_s64:
|
|
; CHECK: {{ext.16.*#8}}
|
|
%xS64x2 = alloca <2 x i64>, align 16
|
|
%__a = alloca <2 x i64>, align 16
|
|
%__b = alloca <2 x i64>, align 16
|
|
%tmp = load <2 x i64>, <2 x i64>* %xS64x2, align 16
|
|
store <2 x i64> %tmp, <2 x i64>* %__a, align 16
|
|
%tmp1 = load <2 x i64>, <2 x i64>* %xS64x2, align 16
|
|
store <2 x i64> %tmp1, <2 x i64>* %__b, align 16
|
|
%tmp2 = load <2 x i64>, <2 x i64>* %__a, align 16
|
|
%tmp3 = bitcast <2 x i64> %tmp2 to <16 x i8>
|
|
%tmp4 = load <2 x i64>, <2 x i64>* %__b, align 16
|
|
%tmp5 = bitcast <2 x i64> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <2 x i64>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <2 x i64>
|
|
%vext = shufflevector <2 x i64> %tmp6, <2 x i64> %tmp7, <2 x i32> <i32 1, i32 2>
|
|
store <2 x i64> %vext, <2 x i64>* %xS64x2, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_vextq_u64() nounwind ssp {
|
|
; CHECK-LABEL: test_vextq_u64:
|
|
; CHECK: {{ext.16.*#8}}
|
|
%xU64x2 = alloca <2 x i64>, align 16
|
|
%__a = alloca <2 x i64>, align 16
|
|
%__b = alloca <2 x i64>, align 16
|
|
%tmp = load <2 x i64>, <2 x i64>* %xU64x2, align 16
|
|
store <2 x i64> %tmp, <2 x i64>* %__a, align 16
|
|
%tmp1 = load <2 x i64>, <2 x i64>* %xU64x2, align 16
|
|
store <2 x i64> %tmp1, <2 x i64>* %__b, align 16
|
|
%tmp2 = load <2 x i64>, <2 x i64>* %__a, align 16
|
|
%tmp3 = bitcast <2 x i64> %tmp2 to <16 x i8>
|
|
%tmp4 = load <2 x i64>, <2 x i64>* %__b, align 16
|
|
%tmp5 = bitcast <2 x i64> %tmp4 to <16 x i8>
|
|
%tmp6 = bitcast <16 x i8> %tmp3 to <2 x i64>
|
|
%tmp7 = bitcast <16 x i8> %tmp5 to <2 x i64>
|
|
%vext = shufflevector <2 x i64> %tmp6, <2 x i64> %tmp7, <2 x i32> <i32 1, i32 2>
|
|
store <2 x i64> %vext, <2 x i64>* %xU64x2, align 16
|
|
ret void
|
|
}
|
|
|
|
; shuffles with an undef second operand can use an EXT also so long as the
|
|
; indices wrap and stay sequential.
|
|
; rdar://12051674
|
|
define <16 x i8> @vext1(<16 x i8> %_a) nounwind {
|
|
; CHECK-LABEL: vext1:
|
|
; CHECK: ext.16b v0, v0, v0, #8
|
|
%vext = shufflevector <16 x i8> %_a, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
ret <16 x i8> %vext
|
|
}
|
|
|
|
; <rdar://problem/12212062>
|
|
define <2 x i64> @vext2(<2 x i64> %p0, <2 x i64> %p1) nounwind readnone ssp {
|
|
entry:
|
|
; CHECK-LABEL: vext2:
|
|
; CHECK: ext.16b v1, v1, v1, #8
|
|
; CHECK: ext.16b v0, v0, v0, #8
|
|
; CHECK: add.2d v0, v0, v1
|
|
%t0 = shufflevector <2 x i64> %p1, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
|
|
%t1 = shufflevector <2 x i64> %p0, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
|
|
%t2 = add <2 x i64> %t1, %t0
|
|
ret <2 x i64> %t2
|
|
}
|