[ARM64-BE] Implement the lane-twiddling logic at AAPCS boundaries for big endian.

The AAPCS states that values passed in registers must have a value as though
they had been loaded with "LDR". LDR is equivalent to "LD1.64 vX.1D" - that is,
loading scalars to vector registers and loading 1-element vectors is equivalent.

The logic implemented here is to ensure that at all call boundaries and during
formal argument lowering all vectors are treated as their bitwidth-based floating
point scalar counterpart, which is always one of f64 or f128 (v2i32 -> f64,
v4i32 -> f128 etc). A BITCAST is inserted so that the appropriate REV will be
generated during code generation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208198 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
James Molloy 2014-05-07 12:33:41 +00:00
parent 1caec99d5d
commit fca7f5c585
3 changed files with 1963 additions and 0 deletions

View File

@ -14,6 +14,9 @@
/// CCIfAlign - Match of the original alignment of the arg
class CCIfAlign<string Align, CCAction A> :
CCIf<!strconcat("ArgFlags.getOrigAlign() == ", Align), A>;
/// CCIfBigEndian - Match only if we're in big endian mode.
class CCIfBigEndian<CCAction A> :
CCIf<"State.getTarget().getDataLayout()->isBigEndian()", A>;
//===----------------------------------------------------------------------===//
// ARM AAPCS64 Calling Convention
@ -23,6 +26,13 @@ def CC_ARM64_AAPCS : CallingConv<[
CCIfType<[v2f32], CCBitConvertToType<v2i32>>,
CCIfType<[v2f64, v4f32], CCBitConvertToType<v2i64>>,
// Big endian vectors must be passed as if they were 1-element vectors so that
// their lanes are in a consistent order.
CCIfBigEndian<CCIfType<[v2i32, v2f32, v4i16, v4f16, v8i8],
CCBitConvertToType<f64>>>,
CCIfBigEndian<CCIfType<[v2i64, v2f64, v4i32, v4f32, v8i16, v8f16, v16i8],
CCBitConvertToType<f128>>>,
// An SRet is passed in X8, not X0 like a normal pointer parameter.
CCIfSRet<CCIfType<[i64], CCAssignToRegWithShadow<[X8], [W8]>>>,
@ -67,6 +77,13 @@ def RetCC_ARM64_AAPCS : CallingConv<[
CCIfType<[v2f32], CCBitConvertToType<v2i32>>,
CCIfType<[v2f64, v4f32], CCBitConvertToType<v2i64>>,
// Big endian vectors must be passed as if they were 1-element vectors so that
// their lanes are in a consistent order.
CCIfBigEndian<CCIfType<[v2i32, v2f32, v4i16, v4f16, v8i8],
CCBitConvertToType<f64>>>,
CCIfBigEndian<CCIfType<[v2i64, v2f64, v4i32, v4f32, v8i16, v8f16, v16i8],
CCBitConvertToType<f128>>>,
CCIfType<[i32], CCAssignToRegWithShadow<[W0, W1, W2, W3, W4, W5, W6, W7],
[X0, X1, X2, X3, X4, X5, X6, X7]>>,
CCIfType<[i64], CCAssignToRegWithShadow<[X0, X1, X2, X3, X4, X5, X6, X7],

View File

@ -0,0 +1,847 @@
; RUN: llc -mtriple arm64_be < %s -arm64-load-store-opt=false -o - | FileCheck %s
; CHECK-LABEL: test_i64_f64:
define i64 @test_i64_f64(double %p) {
; CHECK-NOT: rev
%1 = fadd double %p, %p
%2 = bitcast double %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_i64_v1i64:
define i64 @test_i64_v1i64(<1 x i64> %p) {
; CHECK-NOT: rev
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_i64_v2f32:
define i64 @test_i64_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_i64_v2i32:
define i64 @test_i64_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_i64_v4i16:
define i64 @test_i64_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_i64_v8i8:
define i64 @test_i64_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to i64
%3 = add i64 %2, %2
ret i64 %3
}
; CHECK-LABEL: test_f64_i64:
define double @test_f64_i64(i64 %p) {
; CHECK-NOT: rev
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_f64_v1i64:
define double @test_f64_v1i64(<1 x i64> %p) {
; CHECK-NOT: rev
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_f64_v2f32:
define double @test_f64_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_f64_v2i32:
define double @test_f64_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_f64_v4i16:
define double @test_f64_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_f64_v8i8:
define double @test_f64_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to double
%3 = fadd double %2, %2
ret double %3
}
; CHECK-LABEL: test_v1i64_i64:
define <1 x i64> @test_v1i64_i64(i64 %p) {
; CHECK-NOT: rev
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v1i64_f64:
define <1 x i64> @test_v1i64_f64(double %p) {
; CHECK-NOT: rev
%1 = fadd double %p, %p
%2 = bitcast double %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v1i64_v2f32:
define <1 x i64> @test_v1i64_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v1i64_v2i32:
define <1 x i64> @test_v1i64_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v1i64_v4i16:
define <1 x i64> @test_v1i64_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v1i64_v8i8:
define <1 x i64> @test_v1i64_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to <1 x i64>
%3 = add <1 x i64> %2, %2
ret <1 x i64> %3
}
; CHECK-LABEL: test_v2f32_i64:
define <2 x float> @test_v2f32_i64(i64 %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2f32_f64:
define <2 x float> @test_v2f32_f64(double %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd double %p, %p
%2 = bitcast double %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2f32_v1i64:
define <2 x float> @test_v2f32_v1i64(<1 x i64> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2f32_v2i32:
define <2 x float> @test_v2f32_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2f32_v4i16:
define <2 x float> @test_v2f32_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2f32_v8i8:
define <2 x float> @test_v2f32_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to <2 x float>
%3 = fadd <2 x float> %2, %2
ret <2 x float> %3
}
; CHECK-LABEL: test_v2i32_i64:
define <2 x i32> @test_v2i32_i64(i64 %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v2i32_f64:
define <2 x i32> @test_v2i32_f64(double %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd double %p, %p
%2 = bitcast double %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v2i32_v1i64:
define <2 x i32> @test_v2i32_v1i64(<1 x i64> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v2i32_v2f32:
define <2 x i32> @test_v2i32_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v2i32_v4i16:
define <2 x i32> @test_v2i32_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v2i32_v8i8:
define <2 x i32> @test_v2i32_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
; CHECK: rev64 v{{[0-9]+}}.2s
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to <2 x i32>
%3 = add <2 x i32> %2, %2
ret <2 x i32> %3
}
; CHECK-LABEL: test_v4i16_i64:
define <4 x i16> @test_v4i16_i64(i64 %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v4i16_f64:
define <4 x i16> @test_v4i16_f64(double %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = fadd double %p, %p
%2 = bitcast double %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v4i16_v1i64:
define <4 x i16> @test_v4i16_v1i64(<1 x i64> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v4i16_v2f32:
define <4 x i16> @test_v4i16_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v4i16_v2i32:
define <4 x i16> @test_v4i16_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v4i16_v8i8:
define <4 x i16> @test_v4i16_v8i8(<8 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
; CHECK: rev64 v{{[0-9]+}}.4h
%1 = add <8 x i8> %p, %p
%2 = bitcast <8 x i8> %1 to <4 x i16>
%3 = add <4 x i16> %2, %2
ret <4 x i16> %3
}
; CHECK-LABEL: test_v8i8_i64:
define <8 x i8> @test_v8i8_i64(i64 %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add i64 %p, %p
%2 = bitcast i64 %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_v8i8_f64:
define <8 x i8> @test_v8i8_f64(double %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = fadd double %p, %p
%2 = bitcast double %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_v8i8_v1i64:
define <8 x i8> @test_v8i8_v1i64(<1 x i64> %p) {
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <1 x i64> %p, %p
%2 = bitcast <1 x i64> %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_v8i8_v2f32:
define <8 x i8> @test_v8i8_v2f32(<2 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = fadd <2 x float> %p, %p
%2 = bitcast <2 x float> %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_v8i8_v2i32:
define <8 x i8> @test_v8i8_v2i32(<2 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.2s
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <2 x i32> %p, %p
%2 = bitcast <2 x i32> %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_v8i8_v4i16:
define <8 x i8> @test_v8i8_v4i16(<4 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.4h
; CHECK: rev64 v{{[0-9]+}}.8b
%1 = add <4 x i16> %p, %p
%2 = bitcast <4 x i16> %1 to <8 x i8>
%3 = add <8 x i8> %2, %2
ret <8 x i8> %3
}
; CHECK-LABEL: test_f128_v2f64:
define fp128 @test_f128_v2f64(<2 x double> %p) {
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_f128_v2i64:
define fp128 @test_f128_v2i64(<2 x i64> %p) {
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_f128_v4f32:
define fp128 @test_f128_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_f128_v4i32:
define fp128 @test_f128_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_f128_v8i16:
define fp128 @test_f128_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_f128_v16i8:
define fp128 @test_f128_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to fp128
%3 = fadd fp128 %2, %2
ret fp128 %3
}
; CHECK-LABEL: test_v2f64_f128:
define <2 x double> @test_v2f64_f128(fp128 %p) {
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2f64_v2i64:
define <2 x double> @test_v2f64_v2i64(<2 x i64> %p) {
; CHECK: ext
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2f64_v4f32:
define <2 x double> @test_v2f64_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2f64_v4i32:
define <2 x double> @test_v2f64_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2f64_v8i16:
define <2 x double> @test_v2f64_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2f64_v16i8:
define <2 x double> @test_v2f64_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to <2 x double>
%3 = fadd <2 x double> %2, %2
ret <2 x double> %3
}
; CHECK-LABEL: test_v2i64_f128:
define <2 x i64> @test_v2i64_f128(fp128 %p) {
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v2i64_v2f64:
define <2 x i64> @test_v2i64_v2f64(<2 x double> %p) {
; CHECK: ext
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v2i64_v4f32:
define <2 x i64> @test_v2i64_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v2i64_v4i32:
define <2 x i64> @test_v2i64_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v2i64_v8i16:
define <2 x i64> @test_v2i64_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v2i64_v16i8:
define <2 x i64> @test_v2i64_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to <2 x i64>
%3 = add <2 x i64> %2, %2
ret <2 x i64> %3
}
; CHECK-LABEL: test_v4f32_f128:
define <4 x float> @test_v4f32_f128(fp128 %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4f32_v2f64:
define <4 x float> @test_v4f32_v2f64(<2 x double> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4f32_v2i64:
define <4 x float> @test_v4f32_v2i64(<2 x i64> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4f32_v4i32:
define <4 x float> @test_v4f32_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4f32_v8i16:
define <4 x float> @test_v4f32_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4f32_v16i8:
define <4 x float> @test_v4f32_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to <4 x float>
%3 = fadd <4 x float> %2, %2
ret <4 x float> %3
}
; CHECK-LABEL: test_v4i32_f128:
define <4 x i32> @test_v4i32_f128(fp128 %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v4i32_v2f64:
define <4 x i32> @test_v4i32_v2f64(<2 x double> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v4i32_v2i64:
define <4 x i32> @test_v4i32_v2i64(<2 x i64> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v4i32_v4f32:
define <4 x i32> @test_v4i32_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v4i32_v8i16:
define <4 x i32> @test_v4i32_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v4i32_v16i8:
define <4 x i32> @test_v4i32_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to <4 x i32>
%3 = add <4 x i32> %2, %2
ret <4 x i32> %3
}
; CHECK-LABEL: test_v8i16_f128:
define <8 x i16> @test_v8i16_f128(fp128 %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v8i16_v2f64:
define <8 x i16> @test_v8i16_v2f64(<2 x double> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v8i16_v2i64:
define <8 x i16> @test_v8i16_v2i64(<2 x i64> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v8i16_v4f32:
define <8 x i16> @test_v8i16_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v8i16_v4i32:
define <8 x i16> @test_v8i16_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v8i16_v16i8:
define <8 x i16> @test_v8i16_v16i8(<16 x i8> %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
%1 = add <16 x i8> %p, %p
%2 = bitcast <16 x i8> %1 to <8 x i16>
%3 = add <8 x i16> %2, %2
ret <8 x i16> %3
}
; CHECK-LABEL: test_v16i8_f128:
define <16 x i8> @test_v16i8_f128(fp128 %p) {
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = fadd fp128 %p, %p
%2 = bitcast fp128 %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}
; CHECK-LABEL: test_v16i8_v2f64:
define <16 x i8> @test_v16i8_v2f64(<2 x double> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = fadd <2 x double> %p, %p
%2 = bitcast <2 x double> %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}
; CHECK-LABEL: test_v16i8_v2i64:
define <16 x i8> @test_v16i8_v2i64(<2 x i64> %p) {
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = add <2 x i64> %p, %p
%2 = bitcast <2 x i64> %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}
; CHECK-LABEL: test_v16i8_v4f32:
define <16 x i8> @test_v16i8_v4f32(<4 x float> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = fadd <4 x float> %p, %p
%2 = bitcast <4 x float> %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}
; CHECK-LABEL: test_v16i8_v4i32:
define <16 x i8> @test_v16i8_v4i32(<4 x i32> %p) {
; CHECK: rev64 v{{[0-9]+}}.4s
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = add <4 x i32> %p, %p
%2 = bitcast <4 x i32> %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}
; CHECK-LABEL: test_v16i8_v8i16:
define <16 x i8> @test_v16i8_v8i16(<8 x i16> %p) {
; CHECK: rev64 v{{[0-9]+}}.8h
; CHECK: ext
; CHECK: rev64 v{{[0-9]+}}.16b
; CHECK: ext
%1 = add <8 x i16> %p, %p
%2 = bitcast <8 x i16> %1 to <16 x i8>
%3 = add <16 x i8> %2, %2
ret <16 x i8> %3
}

File diff suppressed because it is too large Load Diff