1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
|
; RUN: llc -march=hexagon -hexagon-bit=0 < %s | FileCheck %s
; This spill should be eliminated.
; CHECK-NOT: vmem(r29+#6)
define void @test(ptr noalias nocapture %key, ptr noalias nocapture %data1) #0 {
entry:
br label %for.body
for.body:
%pkey.0542 = phi ptr [ %key, %entry ], [ null, %for.body ]
%pdata0.0541 = phi ptr [ null, %entry ], [ %add.ptr48, %for.body ]
%pdata1.0540 = phi ptr [ %data1, %entry ], [ %add.ptr49, %for.body ]
%dAccum0.0539 = phi <64 x i32> [ undef, %entry ], [ %84, %for.body ]
%0 = load <32 x i32>, ptr %pkey.0542, align 128
%1 = load <32 x i32>, ptr %pdata0.0541, align 128
%2 = load <32 x i32>, ptr undef, align 128
%arrayidx4 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 2
%3 = load <32 x i32>, ptr %arrayidx4, align 128
%arrayidx5 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 2
%4 = load <32 x i32>, ptr %arrayidx5, align 128
%5 = load <32 x i32>, ptr null, align 128
%6 = load <32 x i32>, ptr undef, align 128
%7 = load <32 x i32>, ptr null, align 128
%arrayidx9 = getelementptr inbounds <32 x i32>, ptr %pkey.0542, i32 3
%arrayidx10 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 6
%8 = load <32 x i32>, ptr %arrayidx10, align 128
%arrayidx12 = getelementptr inbounds <32 x i32>, ptr %pkey.0542, i32 4
%9 = load <32 x i32>, ptr %arrayidx12, align 128
%arrayidx13 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 8
%arrayidx14 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 8
%10 = load <32 x i32>, ptr %arrayidx14, align 128
%arrayidx15 = getelementptr inbounds <32 x i32>, ptr %pkey.0542, i32 5
%11 = load <32 x i32>, ptr %arrayidx15, align 128
%arrayidx16 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 10
%arrayidx17 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 10
%12 = load <32 x i32>, ptr %arrayidx17, align 128
%arrayidx18 = getelementptr inbounds <32 x i32>, ptr %pkey.0542, i32 6
%13 = load <32 x i32>, ptr %arrayidx18, align 128
%arrayidx19 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 12
%14 = load <32 x i32>, ptr %arrayidx19, align 128
%arrayidx20 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 12
%15 = load <32 x i32>, ptr %arrayidx20, align 128
%arrayidx22 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 14
%16 = load <32 x i32>, ptr %arrayidx22, align 128
%arrayidx23 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 14
%17 = load <32 x i32>, ptr %arrayidx23, align 128
%18 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %0, <32 x i32> %9)
%19 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %18, <32 x i32> %9, <32 x i32> %0)
%20 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %18, <32 x i32> %0, <32 x i32> %9)
%21 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %18, <32 x i32> undef, <32 x i32> %1)
%22 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %18, <32 x i32> %10, <32 x i32> undef)
%23 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %5, <32 x i32> %13)
%24 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %13, <32 x i32> %5)
%25 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %5, <32 x i32> %13)
%26 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %14, <32 x i32> %6)
%27 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %6, <32 x i32> %14)
%28 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %15, <32 x i32> %7)
%29 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %23, <32 x i32> %7, <32 x i32> %15)
%30 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %2, <32 x i32> %11)
%31 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> %11, <32 x i32> %2)
%32 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> %2, <32 x i32> %11)
%33 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> undef, <32 x i32> %3)
%34 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> %3, <32 x i32> undef)
%35 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> %12, <32 x i32> %4)
%36 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %30, <32 x i32> %4, <32 x i32> %12)
%37 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> zeroinitializer, <32 x i32> undef)
%38 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> undef, <32 x i32> zeroinitializer)
%39 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> %16, <32 x i32> %8)
%40 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> %8, <32 x i32> %16)
%41 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> %17, <32 x i32> undef)
%42 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> zeroinitializer, <32 x i32> undef, <32 x i32> %17)
%43 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %19, <32 x i32> %24)
%44 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %24, <32 x i32> %19)
%45 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %19, <32 x i32> %24)
%46 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %26, <32 x i32> %21)
%47 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %21, <32 x i32> %26)
%48 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %28, <32 x i32> %22)
%49 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %43, <32 x i32> %22, <32 x i32> %28)
%50 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %20, <32 x i32> %25)
%51 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %50, <32 x i32> %25, <32 x i32> %20)
%52 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %50, <32 x i32> %20, <32 x i32> %25)
%53 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %50, <32 x i32> %27, <32 x i32> undef)
%54 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %50, <32 x i32> undef, <32 x i32> %29)
%55 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %31, <32 x i32> %37)
%56 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %55, <32 x i32> %37, <32 x i32> %31)
%57 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %55, <32 x i32> %31, <32 x i32> %37)
%58 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %55, <32 x i32> %39, <32 x i32> %33)
%59 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %55, <32 x i32> %41, <32 x i32> %35)
%60 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %32, <32 x i32> %38)
%61 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %60, <32 x i32> %40, <32 x i32> %34)
%62 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %60, <32 x i32> %36, <32 x i32> %42)
%63 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %44, <32 x i32> %56)
%64 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %63, <32 x i32> %56, <32 x i32> %44)
%65 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %63, <32 x i32> %58, <32 x i32> %46)
%66 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %63, <32 x i32> %59, <32 x i32> %48)
%67 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %45, <32 x i32> %57)
%68 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %67, <32 x i32> %49, <32 x i32> zeroinitializer)
%69 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %51, <32 x i32> zeroinitializer)
%70 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %69, <32 x i32> %61, <32 x i32> %53)
%71 = tail call <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32> %52, <32 x i32> undef)
%72 = tail call <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1> %71, <32 x i32> %54, <32 x i32> %62)
%73 = tail call <32 x i32> @llvm.hexagon.V6.vshuffeb.128B(<32 x i32> %66, <32 x i32> %65)
%74 = tail call <32 x i32> @llvm.hexagon.V6.vshuffeb.128B(<32 x i32> %68, <32 x i32> undef)
%75 = tail call <32 x i32> @llvm.hexagon.V6.vshuffeb.128B(<32 x i32> zeroinitializer, <32 x i32> %70)
%76 = tail call <32 x i32> @llvm.hexagon.V6.vshuffeb.128B(<32 x i32> %72, <32 x i32> zeroinitializer)
%77 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %dAccum0.0539, <32 x i32> %73, i32 65537)
%78 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %77, <32 x i32> zeroinitializer, i32 65537)
%79 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %78, <32 x i32> zeroinitializer, i32 65537)
%80 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %79, <32 x i32> %74, i32 65537)
%81 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %80, <32 x i32> %75, i32 65537)
%82 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %81, <32 x i32> zeroinitializer, i32 65537)
%83 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %82, <32 x i32> undef, i32 65537)
%84 = tail call <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32> %83, <32 x i32> %76, i32 65537)
store <32 x i32> %64, ptr %pkey.0542, align 128
store <32 x i32> %73, ptr %pdata0.0541, align 128
store <32 x i32> zeroinitializer, ptr %arrayidx4, align 128
store <32 x i32> zeroinitializer, ptr undef, align 128
store <32 x i32> zeroinitializer, ptr %arrayidx20, align 128
store <32 x i32> zeroinitializer, ptr null, align 128
%add.ptr48 = getelementptr inbounds <32 x i32>, ptr %pdata0.0541, i32 16
%add.ptr49 = getelementptr inbounds <32 x i32>, ptr %pdata1.0540, i32 16
br i1 false, label %for.end, label %for.body
for.end:
%85 = tail call <32 x i32> @llvm.hexagon.V6.hi.128B(<64 x i32> %84)
ret void
}
declare <128 x i1> @llvm.hexagon.V6.vgtb.128B(<32 x i32>, <32 x i32>) #1
declare <32 x i32> @llvm.hexagon.V6.vmux.128B(<128 x i1>, <32 x i32>, <32 x i32>) #1
declare <32 x i32> @llvm.hexagon.V6.vshuffeb.128B(<32 x i32>, <32 x i32>) #1
declare <64 x i32> @llvm.hexagon.V6.vmpyuh.acc.128B(<64 x i32>, <32 x i32>, i32) #1
declare <32 x i32> @llvm.hexagon.V6.hi.128B(<64 x i32>) #1
attributes #0 = { nounwind "target-cpu"="hexagonv60" "target-features"="+hvx,+hvx-length128b" }
attributes #1 = { nounwind readnone }
|
