1 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2 ; RUN: opt < %s -mtriple=x86_64-unknown -mattr=+sse2 -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE --check-prefix=SSE2
3 ; RUN: opt < %s -mtriple=x86_64-unknown -mattr=+sse4.2,+popcnt -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE --check-prefix=SSE42
4 ; RUN: opt < %s -mtriple=x86_64-unknown -mattr=+avx,+popcnt -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX1
5 ; RUN: opt < %s -mtriple=x86_64-unknown -mattr=+avx2,+popcnt -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX2
7 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
9 @src64 = common global [4 x i64] zeroinitializer, align 32
10 @dst64 = common global [4 x i64] zeroinitializer, align 32
11 @src32 = common global [8 x i32] zeroinitializer, align 32
12 @dst32 = common global [8 x i32] zeroinitializer, align 32
13 @src16 = common global [16 x i16] zeroinitializer, align 32
14 @dst16 = common global [16 x i16] zeroinitializer, align 32
15 @src8 = common global [32 x i8] zeroinitializer, align 32
16 @dst8 = common global [32 x i8] zeroinitializer, align 32
18 declare i64 @llvm.ctpop.i64(i64)
19 declare i32 @llvm.ctpop.i32(i32)
20 declare i16 @llvm.ctpop.i16(i16)
21 declare i8 @llvm.ctpop.i8(i8)
23 define void @ctpop_2i64() #0 {
24 ; SSE2-LABEL: @ctpop_2i64(
25 ; SSE2-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @src64, align 8
26 ; SSE2-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> [[TMP1]])
27 ; SSE2-NEXT: store <2 x i64> [[TMP2]], ptr @dst64, align 8
30 ; SSE42-LABEL: @ctpop_2i64(
31 ; SSE42-NEXT: [[LD0:%.*]] = load i64, ptr @src64, align 8
32 ; SSE42-NEXT: [[LD1:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i32 0, i64 1), align 8
33 ; SSE42-NEXT: [[CTPOP0:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD0]])
34 ; SSE42-NEXT: [[CTPOP1:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD1]])
35 ; SSE42-NEXT: store i64 [[CTPOP0]], ptr @dst64, align 8
36 ; SSE42-NEXT: store i64 [[CTPOP1]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i32 0, i64 1), align 8
37 ; SSE42-NEXT: ret void
39 ; AVX1-LABEL: @ctpop_2i64(
40 ; AVX1-NEXT: [[LD0:%.*]] = load i64, ptr @src64, align 8
41 ; AVX1-NEXT: [[LD1:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i32 0, i64 1), align 8
42 ; AVX1-NEXT: [[CTPOP0:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD0]])
43 ; AVX1-NEXT: [[CTPOP1:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD1]])
44 ; AVX1-NEXT: store i64 [[CTPOP0]], ptr @dst64, align 8
45 ; AVX1-NEXT: store i64 [[CTPOP1]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i32 0, i64 1), align 8
48 ; AVX2-LABEL: @ctpop_2i64(
49 ; AVX2-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @src64, align 8
50 ; AVX2-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> [[TMP1]])
51 ; AVX2-NEXT: store <2 x i64> [[TMP2]], ptr @dst64, align 8
54 %ld0 = load i64, ptr @src64, align 8
55 %ld1 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i32 0, i64 1), align 8
56 %ctpop0 = call i64 @llvm.ctpop.i64(i64 %ld0)
57 %ctpop1 = call i64 @llvm.ctpop.i64(i64 %ld1)
58 store i64 %ctpop0, ptr @dst64, align 8
59 store i64 %ctpop1, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i32 0, i64 1), align 8
63 define void @ctpop_4i64() #0 {
64 ; SSE2-LABEL: @ctpop_4i64(
65 ; SSE2-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @src64, align 4
66 ; SSE2-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> [[TMP1]])
67 ; SSE2-NEXT: store <2 x i64> [[TMP2]], ptr @dst64, align 4
68 ; SSE2-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
69 ; SSE2-NEXT: [[TMP4:%.*]] = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> [[TMP3]])
70 ; SSE2-NEXT: store <2 x i64> [[TMP4]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
73 ; SSE42-LABEL: @ctpop_4i64(
74 ; SSE42-NEXT: [[LD0:%.*]] = load i64, ptr @src64, align 4
75 ; SSE42-NEXT: [[LD1:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 1), align 4
76 ; SSE42-NEXT: [[LD2:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
77 ; SSE42-NEXT: [[LD3:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 3), align 4
78 ; SSE42-NEXT: [[CTPOP0:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD0]])
79 ; SSE42-NEXT: [[CTPOP1:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD1]])
80 ; SSE42-NEXT: [[CTPOP2:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD2]])
81 ; SSE42-NEXT: [[CTPOP3:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD3]])
82 ; SSE42-NEXT: store i64 [[CTPOP0]], ptr @dst64, align 4
83 ; SSE42-NEXT: store i64 [[CTPOP1]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 1), align 4
84 ; SSE42-NEXT: store i64 [[CTPOP2]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
85 ; SSE42-NEXT: store i64 [[CTPOP3]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 3), align 4
86 ; SSE42-NEXT: ret void
88 ; AVX1-LABEL: @ctpop_4i64(
89 ; AVX1-NEXT: [[LD0:%.*]] = load i64, ptr @src64, align 4
90 ; AVX1-NEXT: [[LD1:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 1), align 4
91 ; AVX1-NEXT: [[LD2:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
92 ; AVX1-NEXT: [[LD3:%.*]] = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 3), align 4
93 ; AVX1-NEXT: [[CTPOP0:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD0]])
94 ; AVX1-NEXT: [[CTPOP1:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD1]])
95 ; AVX1-NEXT: [[CTPOP2:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD2]])
96 ; AVX1-NEXT: [[CTPOP3:%.*]] = call i64 @llvm.ctpop.i64(i64 [[LD3]])
97 ; AVX1-NEXT: store i64 [[CTPOP0]], ptr @dst64, align 4
98 ; AVX1-NEXT: store i64 [[CTPOP1]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 1), align 4
99 ; AVX1-NEXT: store i64 [[CTPOP2]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
100 ; AVX1-NEXT: store i64 [[CTPOP3]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 3), align 4
101 ; AVX1-NEXT: ret void
103 ; AVX2-LABEL: @ctpop_4i64(
104 ; AVX2-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @src64, align 4
105 ; AVX2-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.ctpop.v4i64(<4 x i64> [[TMP1]])
106 ; AVX2-NEXT: store <4 x i64> [[TMP2]], ptr @dst64, align 4
107 ; AVX2-NEXT: ret void
109 %ld0 = load i64, ptr @src64, align 4
110 %ld1 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 1), align 4
111 %ld2 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
112 %ld3 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 3), align 4
113 %ctpop0 = call i64 @llvm.ctpop.i64(i64 %ld0)
114 %ctpop1 = call i64 @llvm.ctpop.i64(i64 %ld1)
115 %ctpop2 = call i64 @llvm.ctpop.i64(i64 %ld2)
116 %ctpop3 = call i64 @llvm.ctpop.i64(i64 %ld3)
117 store i64 %ctpop0, ptr @dst64, align 4
118 store i64 %ctpop1, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 1), align 4
119 store i64 %ctpop2, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
120 store i64 %ctpop3, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 3), align 4
124 define void @ctpop_4i32() #0 {
125 ; SSE2-LABEL: @ctpop_4i32(
126 ; SSE2-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @src32, align 4
127 ; SSE2-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> [[TMP1]])
128 ; SSE2-NEXT: store <4 x i32> [[TMP2]], ptr @dst32, align 4
129 ; SSE2-NEXT: ret void
131 ; SSE42-LABEL: @ctpop_4i32(
132 ; SSE42-NEXT: [[LD0:%.*]] = load i32, ptr @src32, align 4
133 ; SSE42-NEXT: [[LD1:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 4
134 ; SSE42-NEXT: [[LD2:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 4
135 ; SSE42-NEXT: [[LD3:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 4
136 ; SSE42-NEXT: [[CTPOP0:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD0]])
137 ; SSE42-NEXT: [[CTPOP1:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD1]])
138 ; SSE42-NEXT: [[CTPOP2:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD2]])
139 ; SSE42-NEXT: [[CTPOP3:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD3]])
140 ; SSE42-NEXT: store i32 [[CTPOP0]], ptr @dst32, align 4
141 ; SSE42-NEXT: store i32 [[CTPOP1]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 4
142 ; SSE42-NEXT: store i32 [[CTPOP2]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 4
143 ; SSE42-NEXT: store i32 [[CTPOP3]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 4
144 ; SSE42-NEXT: ret void
146 ; AVX-LABEL: @ctpop_4i32(
147 ; AVX-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @src32, align 4
148 ; AVX-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> [[TMP1]])
149 ; AVX-NEXT: store <4 x i32> [[TMP2]], ptr @dst32, align 4
152 %ld0 = load i32, ptr @src32, align 4
153 %ld1 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 4
154 %ld2 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 4
155 %ld3 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 4
156 %ctpop0 = call i32 @llvm.ctpop.i32(i32 %ld0)
157 %ctpop1 = call i32 @llvm.ctpop.i32(i32 %ld1)
158 %ctpop2 = call i32 @llvm.ctpop.i32(i32 %ld2)
159 %ctpop3 = call i32 @llvm.ctpop.i32(i32 %ld3)
160 store i32 %ctpop0, ptr @dst32, align 4
161 store i32 %ctpop1, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 4
162 store i32 %ctpop2, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 4
163 store i32 %ctpop3, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 4
167 define void @ctpop_8i32() #0 {
168 ; SSE2-LABEL: @ctpop_8i32(
169 ; SSE2-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @src32, align 2
170 ; SSE2-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> [[TMP1]])
171 ; SSE2-NEXT: store <4 x i32> [[TMP2]], ptr @dst32, align 2
172 ; SSE2-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 4), align 2
173 ; SSE2-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> [[TMP3]])
174 ; SSE2-NEXT: store <4 x i32> [[TMP4]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 4), align 2
175 ; SSE2-NEXT: ret void
177 ; SSE42-LABEL: @ctpop_8i32(
178 ; SSE42-NEXT: [[LD0:%.*]] = load i32, ptr @src32, align 2
179 ; SSE42-NEXT: [[LD1:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 2
180 ; SSE42-NEXT: [[LD2:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 2
181 ; SSE42-NEXT: [[LD3:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 2
182 ; SSE42-NEXT: [[LD4:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 4), align 2
183 ; SSE42-NEXT: [[LD5:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 5), align 2
184 ; SSE42-NEXT: [[LD6:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 6), align 2
185 ; SSE42-NEXT: [[LD7:%.*]] = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 7), align 2
186 ; SSE42-NEXT: [[CTPOP0:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD0]])
187 ; SSE42-NEXT: [[CTPOP1:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD1]])
188 ; SSE42-NEXT: [[CTPOP2:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD2]])
189 ; SSE42-NEXT: [[CTPOP3:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD3]])
190 ; SSE42-NEXT: [[CTPOP4:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD4]])
191 ; SSE42-NEXT: [[CTPOP5:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD5]])
192 ; SSE42-NEXT: [[CTPOP6:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD6]])
193 ; SSE42-NEXT: [[CTPOP7:%.*]] = call i32 @llvm.ctpop.i32(i32 [[LD7]])
194 ; SSE42-NEXT: store i32 [[CTPOP0]], ptr @dst32, align 2
195 ; SSE42-NEXT: store i32 [[CTPOP1]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 2
196 ; SSE42-NEXT: store i32 [[CTPOP2]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 2
197 ; SSE42-NEXT: store i32 [[CTPOP3]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 2
198 ; SSE42-NEXT: store i32 [[CTPOP4]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 4), align 2
199 ; SSE42-NEXT: store i32 [[CTPOP5]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 5), align 2
200 ; SSE42-NEXT: store i32 [[CTPOP6]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 6), align 2
201 ; SSE42-NEXT: store i32 [[CTPOP7]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 7), align 2
202 ; SSE42-NEXT: ret void
204 ; AVX-LABEL: @ctpop_8i32(
205 ; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, ptr @src32, align 2
206 ; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.ctpop.v8i32(<8 x i32> [[TMP1]])
207 ; AVX-NEXT: store <8 x i32> [[TMP2]], ptr @dst32, align 2
210 %ld0 = load i32, ptr @src32, align 2
211 %ld1 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 2
212 %ld2 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 2
213 %ld3 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 2
214 %ld4 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 4), align 2
215 %ld5 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 5), align 2
216 %ld6 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 6), align 2
217 %ld7 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 7), align 2
218 %ctpop0 = call i32 @llvm.ctpop.i32(i32 %ld0)
219 %ctpop1 = call i32 @llvm.ctpop.i32(i32 %ld1)
220 %ctpop2 = call i32 @llvm.ctpop.i32(i32 %ld2)
221 %ctpop3 = call i32 @llvm.ctpop.i32(i32 %ld3)
222 %ctpop4 = call i32 @llvm.ctpop.i32(i32 %ld4)
223 %ctpop5 = call i32 @llvm.ctpop.i32(i32 %ld5)
224 %ctpop6 = call i32 @llvm.ctpop.i32(i32 %ld6)
225 %ctpop7 = call i32 @llvm.ctpop.i32(i32 %ld7)
226 store i32 %ctpop0, ptr @dst32, align 2
227 store i32 %ctpop1, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 2
228 store i32 %ctpop2, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 2
229 store i32 %ctpop3, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 2
230 store i32 %ctpop4, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 4), align 2
231 store i32 %ctpop5, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 5), align 2
232 store i32 %ctpop6, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 6), align 2
233 store i32 %ctpop7, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 7), align 2
237 define void @ctpop_8i16() #0 {
238 ; CHECK-LABEL: @ctpop_8i16(
239 ; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @src16, align 2
240 ; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> [[TMP1]])
241 ; CHECK-NEXT: store <8 x i16> [[TMP2]], ptr @dst16, align 2
242 ; CHECK-NEXT: ret void
244 %ld0 = load i16, ptr @src16, align 2
245 %ld1 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 1), align 2
246 %ld2 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 2), align 2
247 %ld3 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 3), align 2
248 %ld4 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 4), align 2
249 %ld5 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 5), align 2
250 %ld6 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 6), align 2
251 %ld7 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 7), align 2
252 %ctpop0 = call i16 @llvm.ctpop.i16(i16 %ld0)
253 %ctpop1 = call i16 @llvm.ctpop.i16(i16 %ld1)
254 %ctpop2 = call i16 @llvm.ctpop.i16(i16 %ld2)
255 %ctpop3 = call i16 @llvm.ctpop.i16(i16 %ld3)
256 %ctpop4 = call i16 @llvm.ctpop.i16(i16 %ld4)
257 %ctpop5 = call i16 @llvm.ctpop.i16(i16 %ld5)
258 %ctpop6 = call i16 @llvm.ctpop.i16(i16 %ld6)
259 %ctpop7 = call i16 @llvm.ctpop.i16(i16 %ld7)
260 store i16 %ctpop0, ptr @dst16, align 2
261 store i16 %ctpop1, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 1), align 2
262 store i16 %ctpop2, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 2), align 2
263 store i16 %ctpop3, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 3), align 2
264 store i16 %ctpop4, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 4), align 2
265 store i16 %ctpop5, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 5), align 2
266 store i16 %ctpop6, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 6), align 2
267 store i16 %ctpop7, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 7), align 2
271 define void @ctpop_16i16() #0 {
272 ; SSE-LABEL: @ctpop_16i16(
273 ; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @src16, align 2
274 ; SSE-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> [[TMP1]])
275 ; SSE-NEXT: store <8 x i16> [[TMP2]], ptr @dst16, align 2
276 ; SSE-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 8), align 2
277 ; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> [[TMP3]])
278 ; SSE-NEXT: store <8 x i16> [[TMP4]], ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 8), align 2
281 ; AVX-LABEL: @ctpop_16i16(
282 ; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, ptr @src16, align 2
283 ; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.ctpop.v16i16(<16 x i16> [[TMP1]])
284 ; AVX-NEXT: store <16 x i16> [[TMP2]], ptr @dst16, align 2
287 %ld0 = load i16, ptr @src16, align 2
288 %ld1 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 1), align 2
289 %ld2 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 2), align 2
290 %ld3 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 3), align 2
291 %ld4 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 4), align 2
292 %ld5 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 5), align 2
293 %ld6 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 6), align 2
294 %ld7 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 7), align 2
295 %ld8 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 8), align 2
296 %ld9 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 9), align 2
297 %ld10 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 10), align 2
298 %ld11 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 11), align 2
299 %ld12 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 12), align 2
300 %ld13 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 13), align 2
301 %ld14 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 14), align 2
302 %ld15 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 15), align 2
303 %ctpop0 = call i16 @llvm.ctpop.i16(i16 %ld0)
304 %ctpop1 = call i16 @llvm.ctpop.i16(i16 %ld1)
305 %ctpop2 = call i16 @llvm.ctpop.i16(i16 %ld2)
306 %ctpop3 = call i16 @llvm.ctpop.i16(i16 %ld3)
307 %ctpop4 = call i16 @llvm.ctpop.i16(i16 %ld4)
308 %ctpop5 = call i16 @llvm.ctpop.i16(i16 %ld5)
309 %ctpop6 = call i16 @llvm.ctpop.i16(i16 %ld6)
310 %ctpop7 = call i16 @llvm.ctpop.i16(i16 %ld7)
311 %ctpop8 = call i16 @llvm.ctpop.i16(i16 %ld8)
312 %ctpop9 = call i16 @llvm.ctpop.i16(i16 %ld9)
313 %ctpop10 = call i16 @llvm.ctpop.i16(i16 %ld10)
314 %ctpop11 = call i16 @llvm.ctpop.i16(i16 %ld11)
315 %ctpop12 = call i16 @llvm.ctpop.i16(i16 %ld12)
316 %ctpop13 = call i16 @llvm.ctpop.i16(i16 %ld13)
317 %ctpop14 = call i16 @llvm.ctpop.i16(i16 %ld14)
318 %ctpop15 = call i16 @llvm.ctpop.i16(i16 %ld15)
319 store i16 %ctpop0 , ptr @dst16, align 2
320 store i16 %ctpop1 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 1), align 2
321 store i16 %ctpop2 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 2), align 2
322 store i16 %ctpop3 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 3), align 2
323 store i16 %ctpop4 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 4), align 2
324 store i16 %ctpop5 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 5), align 2
325 store i16 %ctpop6 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 6), align 2
326 store i16 %ctpop7 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 7), align 2
327 store i16 %ctpop8 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 8), align 2
328 store i16 %ctpop9 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 9), align 2
329 store i16 %ctpop10, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 10), align 2
330 store i16 %ctpop11, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 11), align 2
331 store i16 %ctpop12, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 12), align 2
332 store i16 %ctpop13, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 13), align 2
333 store i16 %ctpop14, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 14), align 2
334 store i16 %ctpop15, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 15), align 2
338 define void @ctpop_16i8() #0 {
339 ; CHECK-LABEL: @ctpop_16i8(
340 ; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @src8, align 1
341 ; CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> [[TMP1]])
342 ; CHECK-NEXT: store <16 x i8> [[TMP2]], ptr @dst8, align 1
343 ; CHECK-NEXT: ret void
345 %ld0 = load i8, ptr @src8, align 1
346 %ld1 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 1), align 1
347 %ld2 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 2), align 1
348 %ld3 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 3), align 1
349 %ld4 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 4), align 1
350 %ld5 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 5), align 1
351 %ld6 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 6), align 1
352 %ld7 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 7), align 1
353 %ld8 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 8), align 1
354 %ld9 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 9), align 1
355 %ld10 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 10), align 1
356 %ld11 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 11), align 1
357 %ld12 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 12), align 1
358 %ld13 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 13), align 1
359 %ld14 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 14), align 1
360 %ld15 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 15), align 1
361 %ctpop0 = call i8 @llvm.ctpop.i8(i8 %ld0)
362 %ctpop1 = call i8 @llvm.ctpop.i8(i8 %ld1)
363 %ctpop2 = call i8 @llvm.ctpop.i8(i8 %ld2)
364 %ctpop3 = call i8 @llvm.ctpop.i8(i8 %ld3)
365 %ctpop4 = call i8 @llvm.ctpop.i8(i8 %ld4)
366 %ctpop5 = call i8 @llvm.ctpop.i8(i8 %ld5)
367 %ctpop6 = call i8 @llvm.ctpop.i8(i8 %ld6)
368 %ctpop7 = call i8 @llvm.ctpop.i8(i8 %ld7)
369 %ctpop8 = call i8 @llvm.ctpop.i8(i8 %ld8)
370 %ctpop9 = call i8 @llvm.ctpop.i8(i8 %ld9)
371 %ctpop10 = call i8 @llvm.ctpop.i8(i8 %ld10)
372 %ctpop11 = call i8 @llvm.ctpop.i8(i8 %ld11)
373 %ctpop12 = call i8 @llvm.ctpop.i8(i8 %ld12)
374 %ctpop13 = call i8 @llvm.ctpop.i8(i8 %ld13)
375 %ctpop14 = call i8 @llvm.ctpop.i8(i8 %ld14)
376 %ctpop15 = call i8 @llvm.ctpop.i8(i8 %ld15)
377 store i8 %ctpop0 , ptr @dst8, align 1
378 store i8 %ctpop1 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 1), align 1
379 store i8 %ctpop2 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 2), align 1
380 store i8 %ctpop3 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 3), align 1
381 store i8 %ctpop4 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 4), align 1
382 store i8 %ctpop5 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 5), align 1
383 store i8 %ctpop6 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 6), align 1
384 store i8 %ctpop7 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 7), align 1
385 store i8 %ctpop8 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 8), align 1
386 store i8 %ctpop9 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 9), align 1
387 store i8 %ctpop10, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 10), align 1
388 store i8 %ctpop11, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 11), align 1
389 store i8 %ctpop12, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 12), align 1
390 store i8 %ctpop13, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 13), align 1
391 store i8 %ctpop14, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 14), align 1
392 store i8 %ctpop15, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 15), align 1
396 define void @ctpop_32i8() #0 {
397 ; SSE-LABEL: @ctpop_32i8(
398 ; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @src8, align 1
399 ; SSE-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> [[TMP1]])
400 ; SSE-NEXT: store <16 x i8> [[TMP2]], ptr @dst8, align 1
401 ; SSE-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 16), align 1
402 ; SSE-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> [[TMP3]])
403 ; SSE-NEXT: store <16 x i8> [[TMP4]], ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 16), align 1
406 ; AVX-LABEL: @ctpop_32i8(
407 ; AVX-NEXT: [[TMP1:%.*]] = load <32 x i8>, ptr @src8, align 1
408 ; AVX-NEXT: [[TMP2:%.*]] = call <32 x i8> @llvm.ctpop.v32i8(<32 x i8> [[TMP1]])
409 ; AVX-NEXT: store <32 x i8> [[TMP2]], ptr @dst8, align 1
412 %ld0 = load i8, ptr @src8, align 1
413 %ld1 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 1), align 1
414 %ld2 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 2), align 1
415 %ld3 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 3), align 1
416 %ld4 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 4), align 1
417 %ld5 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 5), align 1
418 %ld6 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 6), align 1
419 %ld7 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 7), align 1
420 %ld8 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 8), align 1
421 %ld9 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 9), align 1
422 %ld10 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 10), align 1
423 %ld11 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 11), align 1
424 %ld12 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 12), align 1
425 %ld13 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 13), align 1
426 %ld14 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 14), align 1
427 %ld15 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 15), align 1
428 %ld16 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 16), align 1
429 %ld17 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 17), align 1
430 %ld18 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 18), align 1
431 %ld19 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 19), align 1
432 %ld20 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 20), align 1
433 %ld21 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 21), align 1
434 %ld22 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 22), align 1
435 %ld23 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 23), align 1
436 %ld24 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 24), align 1
437 %ld25 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 25), align 1
438 %ld26 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 26), align 1
439 %ld27 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 27), align 1
440 %ld28 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 28), align 1
441 %ld29 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 29), align 1
442 %ld30 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 30), align 1
443 %ld31 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 31), align 1
444 %ctpop0 = call i8 @llvm.ctpop.i8(i8 %ld0)
445 %ctpop1 = call i8 @llvm.ctpop.i8(i8 %ld1)
446 %ctpop2 = call i8 @llvm.ctpop.i8(i8 %ld2)
447 %ctpop3 = call i8 @llvm.ctpop.i8(i8 %ld3)
448 %ctpop4 = call i8 @llvm.ctpop.i8(i8 %ld4)
449 %ctpop5 = call i8 @llvm.ctpop.i8(i8 %ld5)
450 %ctpop6 = call i8 @llvm.ctpop.i8(i8 %ld6)
451 %ctpop7 = call i8 @llvm.ctpop.i8(i8 %ld7)
452 %ctpop8 = call i8 @llvm.ctpop.i8(i8 %ld8)
453 %ctpop9 = call i8 @llvm.ctpop.i8(i8 %ld9)
454 %ctpop10 = call i8 @llvm.ctpop.i8(i8 %ld10)
455 %ctpop11 = call i8 @llvm.ctpop.i8(i8 %ld11)
456 %ctpop12 = call i8 @llvm.ctpop.i8(i8 %ld12)
457 %ctpop13 = call i8 @llvm.ctpop.i8(i8 %ld13)
458 %ctpop14 = call i8 @llvm.ctpop.i8(i8 %ld14)
459 %ctpop15 = call i8 @llvm.ctpop.i8(i8 %ld15)
460 %ctpop16 = call i8 @llvm.ctpop.i8(i8 %ld16)
461 %ctpop17 = call i8 @llvm.ctpop.i8(i8 %ld17)
462 %ctpop18 = call i8 @llvm.ctpop.i8(i8 %ld18)
463 %ctpop19 = call i8 @llvm.ctpop.i8(i8 %ld19)
464 %ctpop20 = call i8 @llvm.ctpop.i8(i8 %ld20)
465 %ctpop21 = call i8 @llvm.ctpop.i8(i8 %ld21)
466 %ctpop22 = call i8 @llvm.ctpop.i8(i8 %ld22)
467 %ctpop23 = call i8 @llvm.ctpop.i8(i8 %ld23)
468 %ctpop24 = call i8 @llvm.ctpop.i8(i8 %ld24)
469 %ctpop25 = call i8 @llvm.ctpop.i8(i8 %ld25)
470 %ctpop26 = call i8 @llvm.ctpop.i8(i8 %ld26)
471 %ctpop27 = call i8 @llvm.ctpop.i8(i8 %ld27)
472 %ctpop28 = call i8 @llvm.ctpop.i8(i8 %ld28)
473 %ctpop29 = call i8 @llvm.ctpop.i8(i8 %ld29)
474 %ctpop30 = call i8 @llvm.ctpop.i8(i8 %ld30)
475 %ctpop31 = call i8 @llvm.ctpop.i8(i8 %ld31)
476 store i8 %ctpop0 , ptr @dst8, align 1
477 store i8 %ctpop1 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 1), align 1
478 store i8 %ctpop2 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 2), align 1
479 store i8 %ctpop3 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 3), align 1
480 store i8 %ctpop4 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 4), align 1
481 store i8 %ctpop5 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 5), align 1
482 store i8 %ctpop6 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 6), align 1
483 store i8 %ctpop7 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 7), align 1
484 store i8 %ctpop8 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 8), align 1
485 store i8 %ctpop9 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 9), align 1
486 store i8 %ctpop10, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 10), align 1
487 store i8 %ctpop11, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 11), align 1
488 store i8 %ctpop12, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 12), align 1
489 store i8 %ctpop13, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 13), align 1
490 store i8 %ctpop14, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 14), align 1
491 store i8 %ctpop15, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 15), align 1
492 store i8 %ctpop16, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 16), align 1
493 store i8 %ctpop17, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 17), align 1
494 store i8 %ctpop18, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 18), align 1
495 store i8 %ctpop19, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 19), align 1
496 store i8 %ctpop20, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 20), align 1
497 store i8 %ctpop21, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 21), align 1
498 store i8 %ctpop22, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 22), align 1
499 store i8 %ctpop23, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 23), align 1
500 store i8 %ctpop24, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 24), align 1
501 store i8 %ctpop25, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 25), align 1
502 store i8 %ctpop26, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 26), align 1
503 store i8 %ctpop27, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 27), align 1
504 store i8 %ctpop28, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 28), align 1
505 store i8 %ctpop29, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 29), align 1
506 store i8 %ctpop30, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 30), align 1
507 store i8 %ctpop31, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 31), align 1
511 attributes #0 = { nounwind }