; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=sse | FileCheck %s
define i32 @test_load(ptr %a) {
; CHECK-LABEL: test_load:
; CHECK: ## %bb.0: ## %start
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: movaps (%eax), %xmm0
; CHECK-NEXT: movmskps %xmm0, %eax
; CHECK-NEXT: retl
start:
%0 = load <4 x i32>, ptr %a, align 16
%1 = icmp slt <4 x i32> %0, zeroinitializer
%2 = bitcast <4 x i1> %1 to i4
%3 = zext i4 %2 to i32
ret i32 %3
}
define i32 @test_bitcast(<4 x float> %a) {
; CHECK-LABEL: test_bitcast:
; CHECK: ## %bb.0: ## %start
; CHECK-NEXT: movmskps %xmm0, %eax
; CHECK-NEXT: retl
start:
%0 = bitcast <4 x float> %a to <4 x i32>
%1 = icmp slt <4 x i32> %0, zeroinitializer
%2 = bitcast <4 x i1> %1 to i4
%3 = zext i4 %2 to i32
ret i32 %3
}
define i32 @test_and(<4 x float> %a, <4 x float> %b) {
; CHECK-LABEL: test_and:
; CHECK: ## %bb.0: ## %start
; CHECK-NEXT: andps %xmm1, %xmm0
; CHECK-NEXT: movmskps %xmm0, %eax
; CHECK-NEXT: retl
start:
%0 = bitcast <4 x float> %a to <4 x i32>
%1 = bitcast <4 x float> %b to <4 x i32>
%2 = icmp slt <4 x i32> %0, zeroinitializer
%3 = icmp slt <4 x i32> %1, zeroinitializer
%4 = and <4 x i1> %2, %3
%5 = bitcast <4 x i1> %4 to i4
%6 = zext i4 %5 to i32
ret i32 %6
}
define i32 @test_or(<4 x float> %a, <4 x float> %b) {
; CHECK-LABEL: test_or:
; CHECK: ## %bb.0: ## %start
; CHECK-NEXT: orps %xmm1, %xmm0
; CHECK-NEXT: movmskps %xmm0, %eax
; CHECK-NEXT: retl
start:
%0 = bitcast <4 x float> %a to <4 x i32>
%1 = bitcast <4 x float> %b to <4 x i32>
%2 = icmp slt <4 x i32> %0, zeroinitializer
%3 = icmp slt <4 x i32> %1, zeroinitializer
%4 = or <4 x i1> %2, %3
%5 = bitcast <4 x i1> %4 to i4
%6 = zext i4 %5 to i32
ret i32 %6
}
define i32 @test_xor(<4 x float> %a, <4 x float> %b) {
; CHECK-LABEL: test_xor:
; CHECK: ## %bb.0: ## %start
; CHECK-NEXT: xorps %xmm1, %xmm0
; CHECK-NEXT: movmskps %xmm0, %eax
; CHECK-NEXT: retl
start:
%0 = bitcast <4 x float> %a to <4 x i32>
%1 = bitcast <4 x float> %b to <4 x i32>
%2 = icmp slt <4 x i32> %0, zeroinitializer
%3 = icmp slt <4 x i32> %1, zeroinitializer
%4 = xor <4 x i1> %2, %3
%5 = bitcast <4 x i1> %4 to i4
%6 = zext i4 %5 to i32
ret i32 %6
}