1 ; RUN: llc < %s -mtriple=i386-apple-darwin -mcpu=yonah -regalloc=basic | FileCheck %s
2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -regalloc=basic | FileCheck %s
4 ; This testcase should need to spill the -1 value on both x86-32 and x86-64,
5 ; so it shouldn't use pcmpeqd to materialize an all-ones vector; it
6 ; should use a constant-pool load instead.
8 ; RAGreedy defeats the test by splitting live ranges.
10 ; Constant pool all-ones vector:
11 ; CHECK: .space 16,255
13 ; No pcmpeqd instructions, everybody uses the constant pool.
14 ; CHECK-LABEL: program_1:
17 %struct.__ImageExecInfo = type <{ <4 x i32>, <4 x float>, <2 x i64>, i8*, i8*, i8*, i32, i32, i32, i32, i32 }>
18 %struct._cl_image_format_t = type <{ i32, i32, i32 }>
19 %struct._image2d_t = type <{ i8*, %struct._cl_image_format_t, i32, i32, i32, i32, i32, i32 }>
21 define void @program_1(%struct._image2d_t* %dest, %struct._image2d_t* %t0, <4 x float> %p0, <4 x float> %p1, <4 x float> %p4, <4 x float> %p5, <4 x float> %p6) nounwind {
23 %tmp3.i = load i32, i32* null ; <i32> [#uses=1]
24 %cmp = icmp slt i32 0, %tmp3.i ; <i1> [#uses=1]
25 br i1 %cmp, label %forcond, label %ifthen
27 ifthen: ; preds = %entry
30 forcond: ; preds = %entry
31 %tmp3.i536 = load i32, i32* null ; <i32> [#uses=1]
32 %cmp12 = icmp slt i32 0, %tmp3.i536 ; <i1> [#uses=1]
33 br i1 %cmp12, label %forbody, label %afterfor
35 forbody: ; preds = %forcond
36 %bitcast204.i104 = bitcast <4 x i32> zeroinitializer to <4 x float> ; <<4 x float>> [#uses=1]
37 %tmp78 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> < float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02 >, <4 x float> zeroinitializer) nounwind ; <<4 x float>> [#uses=2]
38 %tmp79 = call <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float> %tmp78) nounwind ; <<4 x i32>> [#uses=1]
39 %tmp80 = call <4 x float> @llvm.x86.sse2.cvtdq2ps(<4 x i32> %tmp79) nounwind ; <<4 x float>> [#uses=1]
40 %sub140.i = fsub <4 x float> %tmp78, %tmp80 ; <<4 x float>> [#uses=2]
41 %mul166.i = fmul <4 x float> zeroinitializer, %sub140.i ; <<4 x float>> [#uses=1]
42 %add167.i = fadd <4 x float> %mul166.i, < float 0x3FE62ACB60000000, float 0x3FE62ACB60000000, float 0x3FE62ACB60000000, float 0x3FE62ACB60000000 > ; <<4 x float>> [#uses=1]
43 %mul171.i = fmul <4 x float> %add167.i, %sub140.i ; <<4 x float>> [#uses=1]
44 %add172.i = fadd <4 x float> %mul171.i, < float 0x3FF0000A40000000, float 0x3FF0000A40000000, float 0x3FF0000A40000000, float 0x3FF0000A40000000 > ; <<4 x float>> [#uses=1]
45 %bitcast176.i = bitcast <4 x float> %add172.i to <4 x i32> ; <<4 x i32>> [#uses=1]
46 %andnps178.i = add <4 x i32> %bitcast176.i, <i32 1, i32 1, i32 1, i32 1> ; <<4 x i32>> [#uses=1]
47 %bitcast179.i = bitcast <4 x i32> %andnps178.i to <4 x float> ; <<4 x float>> [#uses=1]
48 %mul186.i = fmul <4 x float> %bitcast179.i, zeroinitializer ; <<4 x float>> [#uses=1]
49 %bitcast190.i = bitcast <4 x float> %mul186.i to <4 x i32> ; <<4 x i32>> [#uses=1]
50 %andnps192.i = add <4 x i32> %bitcast190.i, <i32 1, i32 1, i32 1, i32 1> ; <<4 x i32>> [#uses=1]
51 %xorps.i = xor <4 x i32> zeroinitializer, < i32 -1, i32 -1, i32 -1, i32 -1 > ; <<4 x i32>> [#uses=1]
52 %orps203.i = add <4 x i32> %andnps192.i, %xorps.i ; <<4 x i32>> [#uses=1]
53 %bitcast204.i = bitcast <4 x i32> %orps203.i to <4 x float> ; <<4 x float>> [#uses=1]
54 %mul310 = fmul <4 x float> %bitcast204.i104, zeroinitializer ; <<4 x float>> [#uses=2]
55 %mul313 = fmul <4 x float> %bitcast204.i, zeroinitializer ; <<4 x float>> [#uses=1]
56 %cmpunord.i11 = call <4 x float> @llvm.x86.sse.cmp.ps(<4 x float> zeroinitializer, <4 x float> zeroinitializer, i8 3) nounwind ; <<4 x float>> [#uses=1]
57 %tmp83 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %mul310, <4 x float> zeroinitializer) nounwind ; <<4 x float>> [#uses=1]
58 %bitcast.i3 = bitcast <4 x float> %mul310 to <4 x i32> ; <<4 x i32>> [#uses=1]
59 %andps.i5 = and <4 x i32> %bitcast.i3, zeroinitializer ; <<4 x i32>> [#uses=1]
61 call void null(<4 x float> %mul313, <4 x float> %cmpunord.i11, <4 x float> %tmp83, <4 x float> zeroinitializer, %struct.__ImageExecInfo* null, <4 x i32> zeroinitializer) nounwind
63 %tmp84 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %mul313, <4 x float> zeroinitializer) nounwind ; <<4 x float>> [#uses=1]
65 %bitcast6.i13 = bitcast <4 x float> %cmpunord.i11 to <4 x i32> ; <<4 x i32>> [#uses=2]
66 %andps.i14 = add <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %bitcast6.i13 ; <<4 x i32>> [#uses=1]
67 %not.i16 = xor <4 x i32> %bitcast6.i13, < i32 -1, i32 -1, i32 -1, i32 -1 > ; <<4 x i32>> [#uses=1]
68 %andnps.i17 = add <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %not.i16 ; <<4 x i32>> [#uses=1]
69 %orps.i18 = or <4 x i32> %andnps.i17, %andps.i14 ; <<4 x i32>> [#uses=1]
70 %bitcast17.i19 = bitcast <4 x i32> %orps.i18 to <4 x float> ; <<4 x float>> [#uses=1]
72 %bitcast11.i6 = bitcast <4 x float> %tmp83 to <4 x i32> ; <<4 x i32>> [#uses=1]
73 %not.i7 = xor <4 x i32> zeroinitializer, < i32 -1, i32 -1, i32 -1, i32 -1 > ; <<4 x i32>> [#uses=1]
74 %andnps.i8 = and <4 x i32> %bitcast11.i6, %not.i7 ; <<4 x i32>> [#uses=1]
75 %orps.i9 = or <4 x i32> %andnps.i8, %andps.i5 ; <<4 x i32>> [#uses=1]
76 %bitcast17.i10 = bitcast <4 x i32> %orps.i9 to <4 x float> ; <<4 x float>> [#uses=1]
78 %bitcast6.i = bitcast <4 x float> zeroinitializer to <4 x i32> ; <<4 x i32>> [#uses=2]
79 %andps.i = and <4 x i32> zeroinitializer, %bitcast6.i ; <<4 x i32>> [#uses=1]
80 %bitcast11.i = bitcast <4 x float> %tmp84 to <4 x i32> ; <<4 x i32>> [#uses=1]
81 %not.i = xor <4 x i32> %bitcast6.i, < i32 -1, i32 -1, i32 -1, i32 -1 > ; <<4 x i32>> [#uses=1]
82 %andnps.i = and <4 x i32> %bitcast11.i, %not.i ; <<4 x i32>> [#uses=1]
83 %orps.i = or <4 x i32> %andnps.i, %andps.i ; <<4 x i32>> [#uses=1]
84 %bitcast17.i = bitcast <4 x i32> %orps.i to <4 x float> ; <<4 x float>> [#uses=1]
85 call void null(<4 x float> %bitcast17.i19, <4 x float> %bitcast17.i10, <4 x float> %bitcast17.i, <4 x float> zeroinitializer, %struct.__ImageExecInfo* null, <4 x i32> zeroinitializer) nounwind
88 afterfor: ; preds = %forcond
92 declare <4 x float> @llvm.x86.sse.cmp.ps(<4 x float>, <4 x float>, i8) nounwind readnone
94 declare <4 x float> @llvm.x86.sse2.cvtdq2ps(<4 x i32>) nounwind readnone
96 declare <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float>) nounwind readnone
98 declare <4 x float> @llvm.x86.sse.min.ps(<4 x float>, <4 x float>) nounwind readnone