Lines Matching full:inbounds
29 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
30 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
33 ; SSE2-NEXT: store double [[CEIL0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
34 ; SSE2-NEXT: store double [[CEIL1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
49 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
50 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
53 store double %ceil0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
54 store double %ceil1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
60 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
61 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
62 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
63 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
68 ; SSE2-NEXT: store double [[CEIL0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
69 ; SSE2-NEXT: store double [[CEIL1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
70 ; SSE2-NEXT: store double [[CEIL2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
71 ; SSE2-NEXT: store double [[CEIL3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
76 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
80 ; SSE41-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
89 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
90 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
91 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
92 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
97 store double %ceil0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
98 store double %ceil1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
99 store double %ceil2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
100 store double %ceil3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
106 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
107 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
108 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
109 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
110 ; SSE2-NEXT: [[LD4:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
111 ; SSE2-NEXT: [[LD5:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
112 ; SSE2-NEXT: [[LD6:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
113 ; SSE2-NEXT: [[LD7:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
122 ; SSE2-NEXT: store double [[CEIL0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
123 ; SSE2-NEXT: store double [[CEIL1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
124 ; SSE2-NEXT: store double [[CEIL2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
125 ; SSE2-NEXT: store double [[CEIL3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
126 ; SSE2-NEXT: store double [[CEIL4]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
127 ; SSE2-NEXT: store double [[CEIL5]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
128 ; SSE2-NEXT: store double [[CEIL6]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
129 ; SSE2-NEXT: store double [[CEIL7]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
134 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
135 ; SSE41-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 8
136 ; SSE41-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 8
142 ; SSE41-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
143 ; SSE41-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 8
144 ; SSE41-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 8
149 ; AVX1-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
153 ; AVX1-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
158 ; AVX2-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
162 ; AVX2-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
171 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
172 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
173 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
174 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
175 %ld4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
176 %ld5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
177 %ld6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
178 %ld7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
187 store double %ceil0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
188 store double %ceil1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
189 store double %ceil2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
190 store double %ceil3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
191 store double %ceil4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
192 store double %ceil5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
193 store double %ceil6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
194 store double %ceil7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
200 inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
201 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
204 ; SSE2-NEXT: store double [[FLOOR0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
205 ; SSE2-NEXT: store double [[FLOOR1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
220 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
221 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
224 store double %floor0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
225 store double %floor1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
231 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
232 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
233 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
234 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
239 ; SSE2-NEXT: store double [[FLOOR0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
240 ; SSE2-NEXT: store double [[FLOOR1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
241 ; SSE2-NEXT: store double [[FLOOR2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
242 ; SSE2-NEXT: store double [[FLOOR3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
247 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
251 ; SSE41-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
260 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
261 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
262 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
263 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
268 store double %floor0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
269 store double %floor1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
270 store double %floor2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
271 store double %floor3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
277 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
278 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
279 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
280 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
281 ; SSE2-NEXT: [[LD4:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
282 ; SSE2-NEXT: [[LD5:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
283 ; SSE2-NEXT: [[LD6:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
284 ; SSE2-NEXT: [[LD7:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
293 ; SSE2-NEXT: store double [[FLOOR0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
294 ; SSE2-NEXT: store double [[FLOOR1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
295 ; SSE2-NEXT: store double [[FLOOR2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
296 ; SSE2-NEXT: store double [[FLOOR3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
297 ; SSE2-NEXT: store double [[FLOOR4]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
298 ; SSE2-NEXT: store double [[FLOOR5]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
299 ; SSE2-NEXT: store double [[FLOOR6]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
300 ; SSE2-NEXT: store double [[FLOOR7]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
305 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
306 ; SSE41-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 8
307 ; SSE41-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 8
313 ; SSE41-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
314 ; SSE41-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 8
315 ; SSE41-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 8
320 ; AVX1-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
324 ; AVX1-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
329 ; AVX2-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
333 ; AVX2-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
342 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
343 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
344 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
345 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
346 %ld4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
347 %ld5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
348 %ld6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
349 %ld7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
358 store double %floor0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
359 store double %floor1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
360 store double %floor2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
361 store double %floor3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
362 store double %floor4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
363 store double %floor5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
364 store double %floor6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
365 store double %floor7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
371 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
372 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
375 ; SSE2-NEXT: store double [[NEARBYINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
376 ; SSE2-NEXT: store double [[NEARBYINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
391 %ld0 = load double, double* getelementptr inbounds
392 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
395 store double %nearbyint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
396 store double %nearbyint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
402 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
403 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
404 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
405 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
410 ; SSE2-NEXT: store double [[NEARBYINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
411 ; SSE2-NEXT: store double [[NEARBYINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
412 ; SSE2-NEXT: store double [[NEARBYINT2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
413 ; SSE2-NEXT: store double [[NEARBYINT3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
418 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
422 ; SSE41-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
431 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
432 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
433 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
434 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
439 store double %nearbyint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
440 store double %nearbyint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
441 store double %nearbyint2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
442 store double %nearbyint3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
448 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
449 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
450 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
451 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
452 ; SSE2-NEXT: [[LD4:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
453 ; SSE2-NEXT: [[LD5:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
454 ; SSE2-NEXT: [[LD6:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
455 ; SSE2-NEXT: [[LD7:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
464 ; SSE2-NEXT: store double [[NEARBYINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
465 ; SSE2-NEXT: store double [[NEARBYINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
466 ; SSE2-NEXT: store double [[NEARBYINT2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
467 ; SSE2-NEXT: store double [[NEARBYINT3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
468 ; SSE2-NEXT: store double [[NEARBYINT4]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
469 ; SSE2-NEXT: store double [[NEARBYINT5]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
470 ; SSE2-NEXT: store double [[NEARBYINT6]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
471 ; SSE2-NEXT: store double [[NEARBYINT7]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
476 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
477 ; SSE41-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 8
478 ; SSE41-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 8
484 ; SSE41-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
485 ; SSE41-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 8
486 ; SSE41-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 8
491 ; AVX1-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
495 ; AVX1-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
500 ; AVX2-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
504 ; AVX2-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
513 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
514 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
515 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
516 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
517 %ld4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
518 %ld5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
519 %ld6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
520 %ld7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
529 store double %nearbyint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
530 store double %nearbyint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
531 store double %nearbyint2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
532 store double %nearbyint3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
533 store double %nearbyint4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
534 store double %nearbyint5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
535 store double %nearbyint6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
536 store double %nearbyint7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
542 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
543 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
546 ; SSE2-NEXT: store double [[RINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
547 ; SSE2-NEXT: store double [[RINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
562 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
563 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
566 store double %rint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
567 store double %rint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
573 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
574 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
575 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
576 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
581 ; SSE2-NEXT: store double [[RINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
582 ; SSE2-NEXT: store double [[RINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
583 ; SSE2-NEXT: store double [[RINT2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
584 ; SSE2-NEXT: store double [[RINT3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
589 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
593 ; SSE41-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
602 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
603 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
604 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
605 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
610 store double %rint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
611 store double %rint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
612 store double %rint2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
613 store double %rint3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
619 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
620 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
621 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
622 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
623 ; SSE2-NEXT: [[LD4:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
624 ; SSE2-NEXT: [[LD5:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
625 ; SSE2-NEXT: [[LD6:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
626 ; SSE2-NEXT: [[LD7:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
635 ; SSE2-NEXT: store double [[RINT0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
636 ; SSE2-NEXT: store double [[RINT1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
637 ; SSE2-NEXT: store double [[RINT2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
638 ; SSE2-NEXT: store double [[RINT3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
639 ; SSE2-NEXT: store double [[RINT4]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
640 ; SSE2-NEXT: store double [[RINT5]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
641 ; SSE2-NEXT: store double [[RINT6]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
642 ; SSE2-NEXT: store double [[RINT7]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
647 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
648 ; SSE41-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 8
649 ; SSE41-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 8
655 ; SSE41-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
656 ; SSE41-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 8
657 ; SSE41-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 8
662 ; AVX1-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
666 ; AVX1-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
671 ; AVX2-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
675 ; AVX2-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
684 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
685 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
686 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
687 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
688 %ld4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
689 %ld5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
690 %ld6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
691 %ld7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
700 store double %rint0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
701 store double %rint1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
702 store double %rint2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
703 store double %rint3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
704 store double %rint4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
705 store double %rint5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
706 store double %rint6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
707 store double %rint7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
713 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
714 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
717 ; SSE2-NEXT: store double [[TRUNC0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
718 ; SSE2-NEXT: store double [[TRUNC1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
733 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
734 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
737 store double %trunc0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
738 store double %trunc1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
744 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
745 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
746 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
747 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
752 ; SSE2-NEXT: store double [[TRUNC0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
753 ; SSE2-NEXT: store double [[TRUNC1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
754 ; SSE2-NEXT: store double [[TRUNC2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
755 ; SSE2-NEXT: store double [[TRUNC3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
760 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
764 ; SSE41-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
773 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
774 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
775 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
776 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
781 store double %trunc0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
782 store double %trunc1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
783 store double %trunc2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
784 store double %trunc3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
790 ; SSE2-NEXT: [[LD0:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
791 ; SSE2-NEXT: [[LD1:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
792 ; SSE2-NEXT: [[LD2:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
793 ; SSE2-NEXT: [[LD3:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
794 ; SSE2-NEXT: [[LD4:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
795 ; SSE2-NEXT: [[LD5:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
796 ; SSE2-NEXT: [[LD6:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
797 ; SSE2-NEXT: [[LD7:%.*]] = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
806 ; SSE2-NEXT: store double [[TRUNC0]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
807 ; SSE2-NEXT: store double [[TRUNC1]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
808 ; SSE2-NEXT: store double [[TRUNC2]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
809 ; SSE2-NEXT: store double [[TRUNC3]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
810 ; SSE2-NEXT: store double [[TRUNC4]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
811 ; SSE2-NEXT: store double [[TRUNC5]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
812 ; SSE2-NEXT: store double [[TRUNC6]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
813 ; SSE2-NEXT: store double [[TRUNC7]], double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
818 ; SSE41-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
819 ; SSE41-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 8
820 ; SSE41-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 8
826 ; SSE41-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
827 ; SSE41-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 8
828 ; SSE41-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 8
833 ; AVX1-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
837 ; AVX1-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
842 ; AVX2-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 8
846 ; AVX2-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 8
855 %ld0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
856 %ld1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
857 %ld2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
858 %ld3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
859 %ld4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 8
860 %ld5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 8
861 %ld6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 8
862 %ld7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 8
871 store double %trunc0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
872 store double %trunc1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
873 store double %trunc2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
874 store double %trunc3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
875 store double %trunc4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 8
876 store double %trunc5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 8
877 store double %trunc6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 8
878 store double %trunc7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 8
884 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
885 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
886 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
887 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
892 ; SSE2-NEXT: store float [[CEIL0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
893 ; SSE2-NEXT: store float [[CEIL1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
894 ; SSE2-NEXT: store float [[CEIL2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
895 ; SSE2-NEXT: store float [[CEIL3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
910 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
911 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
912 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
913 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
918 store float %ceil0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
919 store float %ceil1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
920 store float %ceil2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
921 store float %ceil3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
927 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
928 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
929 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
930 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
931 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
932 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
933 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
934 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
943 ; SSE2-NEXT: store float [[CEIL0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
944 ; SSE2-NEXT: store float [[CEIL1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
945 ; SSE2-NEXT: store float [[CEIL2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
946 ; SSE2-NEXT: store float [[CEIL3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
947 ; SSE2-NEXT: store float [[CEIL4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
948 ; SSE2-NEXT: store float [[CEIL5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
949 ; SSE2-NEXT: store float [[CEIL6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
950 ; SSE2-NEXT: store float [[CEIL7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
955 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
959 ; SSE41-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
968 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
969 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
970 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
971 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
972 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
973 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
974 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
975 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
984 store float %ceil0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
985 store float %ceil1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
986 store float %ceil2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
987 store float %ceil3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
988 store float %ceil4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
989 store float %ceil5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
990 store float %ceil6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
991 store float %ceil7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
997 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
998 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
999 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1000 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1001 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1002 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1003 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1004 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1005 ; SSE2-NEXT: [[LD8:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
1006 ; SSE2-NEXT: [[LD9:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
1007 ; SSE2-NEXT: [[LD10:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1008 ; SSE2-NEXT: [[LD11:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1009 ; SSE2-NEXT: [[LD12:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1010 ; SSE2-NEXT: [[LD13:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1011 ; SSE2-NEXT: [[LD14:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1012 ; SSE2-NEXT: [[LD15:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1029 ; SSE2-NEXT: store float [[CEIL0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1030 ; SSE2-NEXT: store float [[CEIL1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1031 ; SSE2-NEXT: store float [[CEIL2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1032 ; SSE2-NEXT: store float [[CEIL3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1033 ; SSE2-NEXT: store float [[CEIL4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1034 ; SSE2-NEXT: store float [[CEIL5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1035 ; SSE2-NEXT: store float [[CEIL6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1036 ; SSE2-NEXT: store float [[CEIL7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1037 ; SSE2-NEXT: store float [[CEIL8]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
1038 ; SSE2-NEXT: store float [[CEIL9]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
1039 ; SSE2-NEXT: store float [[CEIL10]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1040 ; SSE2-NEXT: store float [[CEIL11]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1041 ; SSE2-NEXT: store float [[CEIL12]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1042 ; SSE2-NEXT: store float [[CEIL13]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1043 ; SSE2-NEXT: store float [[CEIL14]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1044 ; SSE2-NEXT: store float [[CEIL15]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1049 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1050 ; SSE41-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
1051 ; SSE41-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
1057 ; SSE41-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1058 ; SSE41-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
1059 ; SSE41-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
1064 ; AVX1-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1068 ; AVX1-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1073 ; AVX2-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1077 ; AVX2-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1086 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0 ), align 4
1087 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1 ), align 4
1088 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2 ), align 4
1089 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3 ), align 4
1090 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4 ), align 4
1091 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5 ), align 4
1092 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6 ), align 4
1093 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7 ), align 4
1094 %ld8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8 ), align 4
1095 %ld9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9 ), align 4
1096 %ld10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1097 %ld11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1098 %ld12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1099 %ld13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1100 %ld14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1101 %ld15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1118 store float %ceil0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0 ), align 4
1119 store float %ceil1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1 ), align 4
1120 store float %ceil2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2 ), align 4
1121 store float %ceil3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3 ), align 4
1122 store float %ceil4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4 ), align 4
1123 store float %ceil5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5 ), align 4
1124 store float %ceil6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6 ), align 4
1125 store float %ceil7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7 ), align 4
1126 store float %ceil8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8 ), align 4
1127 store float %ceil9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9 ), align 4
1128 store float %ceil10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1129 inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1130 store float %ceil12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1131 store float %ceil13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1132 store float %ceil14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1133 store float %ceil15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1139 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1140 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1141 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1142 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1147 ; SSE2-NEXT: store float [[FLOOR0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1148 ; SSE2-NEXT: store float [[FLOOR1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1149 ; SSE2-NEXT: store float [[FLOOR2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1150 ; SSE2-NEXT: store float [[FLOOR3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1165 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1166 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1167 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1168 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1173 store float %floor0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1174 store float %floor1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1175 store float %floor2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1176 store float %floor3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1182 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1183 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1184 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1185 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1186 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1187 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1188 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1189 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1198 ; SSE2-NEXT: store float [[FLOOR0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1199 ; SSE2-NEXT: store float [[FLOOR1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1200 ; SSE2-NEXT: store float [[FLOOR2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1201 ; SSE2-NEXT: store float [[FLOOR3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1202 ; SSE2-NEXT: store float [[FLOOR4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1203 ; SSE2-NEXT: store float [[FLOOR5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1204 ; SSE2-NEXT: store float [[FLOOR6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1205 ; SSE2-NEXT: store float [[FLOOR7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1210 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1214 ; SSE41-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1223 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1224 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1225 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1226 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1227 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1228 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1229 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1230 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1239 store float %floor0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1240 store float %floor1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1241 store float %floor2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1242 store float %floor3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1243 store float %floor4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1244 store float %floor5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1245 store float %floor6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1246 store float %floor7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1252 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1253 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1254 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1255 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1256 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1257 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1258 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1259 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1260 ; SSE2-NEXT: [[LD8:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
1261 ; SSE2-NEXT: [[LD9:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
1262 ; SSE2-NEXT: [[LD10:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1263 ; SSE2-NEXT: [[LD11:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1264 ; SSE2-NEXT: [[LD12:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1265 ; SSE2-NEXT: [[LD13:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1266 ; SSE2-NEXT: [[LD14:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1267 ; SSE2-NEXT: [[LD15:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1284 ; SSE2-NEXT: store float [[FLOOR0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1285 ; SSE2-NEXT: store float [[FLOOR1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1286 ; SSE2-NEXT: store float [[FLOOR2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1287 ; SSE2-NEXT: store float [[FLOOR3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1288 ; SSE2-NEXT: store float [[FLOOR4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1289 ; SSE2-NEXT: store float [[FLOOR5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1290 ; SSE2-NEXT: store float [[FLOOR6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1291 ; SSE2-NEXT: store float [[FLOOR7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1292 ; SSE2-NEXT: store float [[FLOOR8]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
1293 ; SSE2-NEXT: store float [[FLOOR9]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
1294 ; SSE2-NEXT: store float [[FLOOR10]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1295 ; SSE2-NEXT: store float [[FLOOR11]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1296 ; SSE2-NEXT: store float [[FLOOR12]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1297 ; SSE2-NEXT: store float [[FLOOR13]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1298 ; SSE2-NEXT: store float [[FLOOR14]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1299 ; SSE2-NEXT: store float [[FLOOR15]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1304 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1305 ; SSE41-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
1306 ; SSE41-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
1312 ; SSE41-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1313 ; SSE41-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
1314 ; SSE41-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
1319 ; AVX1-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1323 ; AVX1-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1328 ; AVX2-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1332 ; AVX2-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1341 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0 ), align 4
1342 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1 ), align 4
1343 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2 ), align 4
1344 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3 ), align 4
1345 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4 ), align 4
1346 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5 ), align 4
1347 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6 ), align 4
1348 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7 ), align 4
1349 %ld8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8 ), align 4
1350 %ld9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9 ), align 4
1351 %ld10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1352 %ld11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1353 %ld12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1354 %ld13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1355 %ld14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1356 %ld15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1373 store float %floor0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0 ), align 4
1374 store float %floor1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1 ), align 4
1375 store float %floor2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2 ), align 4
1376 store float %floor3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3 ), align 4
1377 store float %floor4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4 ), align 4
1378 store float %floor5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5 ), align 4
1379 store float %floor6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6 ), align 4
1380 store float %floor7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7 ), align 4
1381 store float %floor8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8 ), align 4
1382 store float %floor9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9 ), align 4
1383 store float %floor10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1384 store float %floor11, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1385 store float %floor12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1386 store float %floor13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1387 store float %floor14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1388 store float %floor15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1394 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1395 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1396 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1397 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1402 ; SSE2-NEXT: store float [[NEARBYINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1403 ; SSE2-NEXT: store float [[NEARBYINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1404 ; SSE2-NEXT: store float [[NEARBYINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1405 ; SSE2-NEXT: store float [[NEARBYINT3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1420 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1421 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1422 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1423 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1428 store float %nearbyint0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1429 store float %nearbyint1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1430 store float %nearbyint2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1431 store float %nearbyint3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1437 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1438 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1439 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1440 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1441 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1442 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1443 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1444 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1453 ; SSE2-NEXT: store float [[NEARBYINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1454 ; SSE2-NEXT: store float [[NEARBYINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1455 ; SSE2-NEXT: store float [[NEARBYINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1456 ; SSE2-NEXT: store float [[NEARBYINT3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1457 ; SSE2-NEXT: store float [[NEARBYINT4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1458 ; SSE2-NEXT: store float [[NEARBYINT5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1459 ; SSE2-NEXT: store float [[NEARBYINT6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1460 ; SSE2-NEXT: store float [[NEARBYINT7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1465 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1469 ; SSE41-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1478 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1479 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1480 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1481 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1482 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1483 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1484 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1485 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1494 store float %nearbyint0, float* getelementptr inbounds
1495 store float %nearbyint1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1496 store float %nearbyint2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1497 store float %nearbyint3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1498 store float %nearbyint4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1499 store float %nearbyint5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1500 store float %nearbyint6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1501 store float %nearbyint7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1507 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1508 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1509 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1510 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1511 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1512 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1513 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1514 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1515 ; SSE2-NEXT: [[LD8:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
1516 ; SSE2-NEXT: [[LD9:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
1517 ; SSE2-NEXT: [[LD10:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1518 ; SSE2-NEXT: [[LD11:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1519 ; SSE2-NEXT: [[LD12:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1520 ; SSE2-NEXT: [[LD13:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1521 ; SSE2-NEXT: [[LD14:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1522 ; SSE2-NEXT: [[LD15:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1539 ; SSE2-NEXT: store float [[NEARBYINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1540 ; SSE2-NEXT: store float [[NEARBYINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1541 ; SSE2-NEXT: store float [[NEARBYINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1542 ; SSE2-NEXT: store float [[NEARBYINT3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1543 ; SSE2-NEXT: store float [[NEARBYINT4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1544 ; SSE2-NEXT: store float [[NEARBYINT5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1545 ; SSE2-NEXT: store float [[NEARBYINT6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1546 ; SSE2-NEXT: store float [[NEARBYINT7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1547 ; SSE2-NEXT: store float [[NEARBYINT8]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
1548 ; SSE2-NEXT: store float [[NEARBYINT9]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
1549 ; SSE2-NEXT: store float [[NEARBYINT10]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1550 ; SSE2-NEXT: store float [[NEARBYINT11]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1551 ; SSE2-NEXT: store float [[NEARBYINT12]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1552 ; SSE2-NEXT: store float [[NEARBYINT13]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1553 ; SSE2-NEXT: store float [[NEARBYINT14]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1554 ; SSE2-NEXT: store float [[NEARBYINT15]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1559 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1560 ; SSE41-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
1561 ; SSE41-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
1567 ; SSE41-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1568 ; SSE41-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
1569 ; SSE41-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
1574 ; AVX1-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1578 ; AVX1-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1583 ; AVX2-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1587 ; AVX2-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1596 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0 ), align 4
1597 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1 ), align 4
1598 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2 ), align 4
1599 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3 ), align 4
1600 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4 ), align 4
1601 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5 ), align 4
1602 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6 ), align 4
1603 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7 ), align 4
1604 %ld8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8 ), align 4
1605 %ld9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9 ), align 4
1606 %ld10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1607 %ld11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1608 %ld12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1609 %ld13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1610 %ld14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1611 %ld15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1628 store float %nearbyint0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0 ), align 4
1629 store float %nearbyint1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1 ), align 4
1630 store float %nearbyint2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2 ), align 4
1631 store float %nearbyint3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3 ), align 4
1632 store float %nearbyint4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4 ), align 4
1633 store float %nearbyint5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5 ), align 4
1634 store float %nearbyint6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6 ), align 4
1635 store float %nearbyint7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7 ), align 4
1636 store float %nearbyint8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8 ), align 4
1637 store float %nearbyint9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9 ), align 4
1638 store float %nearbyint10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1639 store float %nearbyint11, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1640 store float %nearbyint12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1641 store float %nearbyint13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1642 store float %nearbyint14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1643 store float %nearbyint15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1649 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1650 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1651 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1652 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1657 ; SSE2-NEXT: store float [[RINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1658 ; SSE2-NEXT: store float [[RINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1659 ; SSE2-NEXT: store float [[RINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1660 ; SSE2-NEXT: store float [[RINT3]], float* getelementptr inbounds
1675 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1676 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1677 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1678 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1683 store float %rint0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1684 store float %rint1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1685 store float %rint2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1686 store float %rint3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1692 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1693 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1694 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1695 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1696 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1697 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1698 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1699 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1708 ; SSE2-NEXT: store float [[RINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1709 ; SSE2-NEXT: store float [[RINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1710 ; SSE2-NEXT: store float [[RINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1711 ; SSE2-NEXT: store float [[RINT3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1712 ; SSE2-NEXT: store float [[RINT4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1713 ; SSE2-NEXT: store float [[RINT5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1714 ; SSE2-NEXT: store float [[RINT6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1715 ; SSE2-NEXT: store float [[RINT7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1720 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1724 ; SSE41-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1733 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1734 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1735 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1736 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1737 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1738 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1739 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1740 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1749 store float %rint0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1750 store float %rint1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1751 store float %rint2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1752 store float %rint3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1753 store float %rint4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1754 store float %rint5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1755 store float %rint6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1756 store float %rint7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1762 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1763 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1764 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1765 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1766 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1767 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1768 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1769 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1770 ; SSE2-NEXT: [[LD8:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
1771 ; SSE2-NEXT: [[LD9:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
1772 ; SSE2-NEXT: [[LD10:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1773 ; SSE2-NEXT: [[LD11:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1774 ; SSE2-NEXT: [[LD12:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1775 ; SSE2-NEXT: [[LD13:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1776 ; SSE2-NEXT: [[LD14:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1777 ; SSE2-NEXT: [[LD15:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1794 ; SSE2-NEXT: store float [[RINT0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1795 ; SSE2-NEXT: store float [[RINT1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1796 ; SSE2-NEXT: store float [[RINT2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1797 ; SSE2-NEXT: store float [[RINT3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1798 ; SSE2-NEXT: store float [[RINT4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1799 ; SSE2-NEXT: store float [[RINT5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1800 ; SSE2-NEXT: store float [[RINT6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1801 ; SSE2-NEXT: store float [[RINT7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1802 ; SSE2-NEXT: store float [[RINT8]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
1803 ; SSE2-NEXT: store float [[RINT9]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
1804 ; SSE2-NEXT: store float [[RINT10]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1805 ; SSE2-NEXT: store float [[RINT11]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1806 ; SSE2-NEXT: store float [[RINT12]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1807 ; SSE2-NEXT: store float [[RINT13]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1808 ; SSE2-NEXT: store float [[RINT14]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1809 ; SSE2-NEXT: store float [[RINT15]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1814 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1815 ; SSE41-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
1816 ; SSE41-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
1822 ; SSE41-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1823 ; SSE41-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
1824 ; SSE41-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
1829 ; AVX1-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1833 ; AVX1-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1838 ; AVX2-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
1842 ; AVX2-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
1851 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0 ), align 4
1852 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1 ), align 4
1853 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2 ), align 4
1854 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3 ), align 4
1855 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4 ), align 4
1856 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5 ), align 4
1857 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6 ), align 4
1858 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7 ), align 4
1859 %ld8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8 ), align 4
1860 %ld9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9 ), align 4
1861 %ld10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
1862 %ld11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
1863 %ld12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
1864 %ld13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
1865 %ld14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
1866 %ld15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
1883 store float %rint0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0 ), align 4
1884 store float %rint1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1 ), align 4
1885 store float %rint2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2 ), align 4
1886 store float %rint3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3 ), align 4
1887 store float %rint4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4 ), align 4
1888 store float %rint5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5 ), align 4
1889 store float %rint6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6 ), align 4
1890 store float %rint7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7 ), align 4
1891 store float %rint8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8 ), align 4
1892 store float %rint9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9 ), align 4
1893 store float %rint10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
1894 store float %rint11, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
1895 store float %rint12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
1896 store float %rint13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
1897 store float %rint14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
1898 store float %rint15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
1904 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1905 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1906 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1907 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1912 ; SSE2-NEXT: store float [[TRUNC0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1913 ; SSE2-NEXT: store float [[TRUNC1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1914 ; SSE2-NEXT: store float [[TRUNC2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1915 ; SSE2-NEXT: store float [[TRUNC3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1930 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1931 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1932 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1933 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1938 store float %trunc0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1939 store float %trunc1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1940 store float %trunc2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1941 store float %trunc3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1947 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1948 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1949 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1950 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1951 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1952 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1953 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1954 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
1963 ; SSE2-NEXT: store float [[TRUNC0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
1964 ; SSE2-NEXT: store float [[TRUNC1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
1965 ; SSE2-NEXT: store float [[TRUNC2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
1966 ; SSE2-NEXT: store float [[TRUNC3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
1967 ; SSE2-NEXT: store float [[TRUNC4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
1968 ; SSE2-NEXT: store float [[TRUNC5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
1969 ; SSE2-NEXT: store float [[TRUNC6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
1970 ; SSE2-NEXT: store float [[TRUNC7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
1975 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
1979 ; SSE41-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
1988 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
1989 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
1990 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
1991 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
1992 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
1993 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
1994 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
1995 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
2004 store float %trunc0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
2005 store float %trunc1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
2006 store float %trunc2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
2007 store float %trunc3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
2008 store float %trunc4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
2009 store float %trunc5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
2010 store float %trunc6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
2011 store float %trunc7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
2017 ; SSE2-NEXT: [[LD0:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
2018 ; SSE2-NEXT: [[LD1:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
2019 ; SSE2-NEXT: [[LD2:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
2020 ; SSE2-NEXT: [[LD3:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
2021 ; SSE2-NEXT: [[LD4:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
2022 ; SSE2-NEXT: [[LD5:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
2023 ; SSE2-NEXT: [[LD6:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
2024 ; SSE2-NEXT: [[LD7:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
2025 ; SSE2-NEXT: [[LD8:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
2026 ; SSE2-NEXT: [[LD9:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
2027 ; SSE2-NEXT: [[LD10:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
2028 ; SSE2-NEXT: [[LD11:%.*]] = load float, float* getelementptr inbounds
2029 ; SSE2-NEXT: [[LD12:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
2030 ; SSE2-NEXT: [[LD13:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
2031 ; SSE2-NEXT: [[LD14:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
2032 ; SSE2-NEXT: [[LD15:%.*]] = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
2049 ; SSE2-NEXT: store float [[TRUNC0]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
2050 ; SSE2-NEXT: store float [[TRUNC1]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
2051 ; SSE2-NEXT: store float [[TRUNC2]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
2052 ; SSE2-NEXT: store float [[TRUNC3]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
2053 ; SSE2-NEXT: store float [[TRUNC4]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
2054 ; SSE2-NEXT: store float [[TRUNC5]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
2055 ; SSE2-NEXT: store float [[TRUNC6]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
2056 ; SSE2-NEXT: store float [[TRUNC7]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
2057 ; SSE2-NEXT: store float [[TRUNC8]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
2058 ; SSE2-NEXT: store float [[TRUNC9]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
2059 ; SSE2-NEXT: store float [[TRUNC10]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
2060 ; SSE2-NEXT: store float [[TRUNC11]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
2061 ; SSE2-NEXT: store float [[TRUNC12]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
2062 ; SSE2-NEXT: store float [[TRUNC13]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
2063 ; SSE2-NEXT: store float [[TRUNC14]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
2064 ; SSE2-NEXT: store float [[TRUNC15]], float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
2069 ; SSE41-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
2070 ; SSE41-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
2071 ; SSE41-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
2077 ; SSE41-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
2078 ; SSE41-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
2079 ; SSE41-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
2084 ; AVX1-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
2088 ; AVX1-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
2093 ; AVX2-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
2097 ; AVX2-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
2106 %ld0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0 ), align 4
2107 %ld1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1 ), align 4
2108 %ld2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2 ), align 4
2109 %ld3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3 ), align 4
2110 %ld4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4 ), align 4
2111 %ld5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5 ), align 4
2112 %ld6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6 ), align 4
2113 %ld7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7 ), align 4
2114 %ld8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8 ), align 4
2115 %ld9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9 ), align 4
2116 %ld10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
2117 %ld11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
2118 %ld12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
2119 %ld13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
2120 %ld14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
2121 %ld15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
2138 store float %trunc0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0 ), align 4
2139 store float %trunc1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1 ), align 4
2140 store float %trunc2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2 ), align 4
2141 store float %trunc3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3 ), align 4
2142 store float %trunc4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4 ), align 4
2143 store float %trunc5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5 ), align 4
2144 store float %trunc6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6 ), align 4
2145 store float %trunc7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7 ), align 4
2146 store float %trunc8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8 ), align 4
2147 store float %trunc9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9 ), align 4
2148 store float %trunc10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
2149 store float %trunc11, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
2150 store float %trunc12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
2151 store float %trunc13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
2152 store float %trunc14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
2153 store float %trunc15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
