Lines Matching refs:SSE2

2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=SSE2
9 ; SSE2-LABEL: roundeven_f16:
10 ; SSE2:       ## %bb.0: ## %entry
11 ; SSE2-NEXT:    pushq %rax
12 ; SSE2-NEXT:    .cfi_def_cfa_offset 16
13 ; SSE2-NEXT:    callq ___extendhfsf2
14 ; SSE2-NEXT:    callq _roundevenf
15 ; SSE2-NEXT:    callq ___truncsfhf2
16 ; SSE2-NEXT:    popq %rax
17 ; SSE2-NEXT:    retq
61 ; SSE2-LABEL: roundeven_f32:
62 ; SSE2:       ## %bb.0:
63 ; SSE2-NEXT:    jmp _roundevenf ## TAILCALL
79 ; SSE2-LABEL: roundeven_f64:
80 ; SSE2:       ## %bb.0:
81 ; SSE2-NEXT:    jmp _roundeven ## TAILCALL
97 ; SSE2-LABEL: roundeven_v4f32:
98 ; SSE2:       ## %bb.0:
99 ; SSE2-NEXT:    subq $56, %rsp
100 ; SSE2-NEXT:    .cfi_def_cfa_offset 64
101 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
102 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
103 ; SSE2-NEXT:    callq _roundevenf
104 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
105 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
106 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
107 ; SSE2-NEXT:    callq _roundevenf
108 ; SSE2-NEXT:    unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload
109 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
110 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
111 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
112 ; SSE2-NEXT:    callq _roundevenf
113 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
114 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
115 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
116 ; SSE2-NEXT:    callq _roundevenf
117 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
118 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
119 ; SSE2-NEXT:    unpcklpd (%rsp), %xmm1 ## 16-byte Folded Reload
120 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
121 ; SSE2-NEXT:    movaps %xmm1, %xmm0
122 ; SSE2-NEXT:    addq $56, %rsp
123 ; SSE2-NEXT:    retq
139 ; SSE2-LABEL: roundeven_v2f64:
140 ; SSE2:       ## %bb.0:
141 ; SSE2-NEXT:    subq $40, %rsp
142 ; SSE2-NEXT:    .cfi_def_cfa_offset 48
143 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
144 ; SSE2-NEXT:    callq _roundeven
145 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
146 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
147 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
148 ; SSE2-NEXT:    callq _roundeven
149 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
150 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
151 ; SSE2-NEXT:    movaps %xmm1, %xmm0
152 ; SSE2-NEXT:    addq $40, %rsp
153 ; SSE2-NEXT:    retq
169 ; SSE2-LABEL: roundeven_v8f32:
170 ; SSE2:       ## %bb.0:
171 ; SSE2-NEXT:    subq $72, %rsp
172 ; SSE2-NEXT:    .cfi_def_cfa_offset 80
173 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
174 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
175 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
176 ; SSE2-NEXT:    callq _roundevenf
177 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
178 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
179 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
180 ; SSE2-NEXT:    callq _roundevenf
181 ; SSE2-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
182 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
183 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
184 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
185 ; SSE2-NEXT:    callq _roundevenf
186 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
187 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
188 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
189 ; SSE2-NEXT:    callq _roundevenf
190 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
191 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
192 ; SSE2-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload
193 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
194 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
195 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
196 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
197 ; SSE2-NEXT:    callq _roundevenf
198 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
199 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
200 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
201 ; SSE2-NEXT:    callq _roundevenf
202 ; SSE2-NEXT:    unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload
203 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
204 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
205 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
206 ; SSE2-NEXT:    callq _roundevenf
207 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
208 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
209 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
210 ; SSE2-NEXT:    callq _roundevenf
211 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
212 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
213 ; SSE2-NEXT:    unpcklpd (%rsp), %xmm1 ## 16-byte Folded Reload
214 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
215 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
216 ; SSE2-NEXT:    addq $72, %rsp
217 ; SSE2-NEXT:    retq
234 ; SSE2-LABEL: roundeven_v4f64:
235 ; SSE2:       ## %bb.0:
236 ; SSE2-NEXT:    subq $56, %rsp
237 ; SSE2-NEXT:    .cfi_def_cfa_offset 64
238 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
239 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
240 ; SSE2-NEXT:    callq _roundeven
241 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
242 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
243 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
244 ; SSE2-NEXT:    callq _roundeven
245 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
246 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
247 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
248 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
249 ; SSE2-NEXT:    callq _roundeven
250 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
251 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
252 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
253 ; SSE2-NEXT:    callq _roundeven
254 ; SSE2-NEXT:    movaps (%rsp), %xmm1 ## 16-byte Reload
255 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
256 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
257 ; SSE2-NEXT:    addq $56, %rsp
258 ; SSE2-NEXT:    retq
275 ; SSE2-LABEL: roundeven_v16f32:
276 ; SSE2:       ## %bb.0:
277 ; SSE2-NEXT:    subq $104, %rsp
278 ; SSE2-NEXT:    .cfi_def_cfa_offset 112
279 ; SSE2-NEXT:    movaps %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
280 ; SSE2-NEXT:    movaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
281 ; SSE2-NEXT:    movaps %xmm1, (%rsp) ## 16-byte Spill
282 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
283 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
284 ; SSE2-NEXT:    callq _roundevenf
285 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
286 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
287 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
288 ; SSE2-NEXT:    callq _roundevenf
289 ; SSE2-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
290 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
291 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
292 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
293 ; SSE2-NEXT:    callq _roundevenf
294 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
295 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
296 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
297 ; SSE2-NEXT:    callq _roundevenf
298 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
299 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
300 ; SSE2-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload
301 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
302 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
303 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
304 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
305 ; SSE2-NEXT:    callq _roundevenf
306 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
307 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
308 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
309 ; SSE2-NEXT:    callq _roundevenf
310 ; SSE2-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
311 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
312 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
313 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
314 ; SSE2-NEXT:    callq _roundevenf
315 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
316 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
317 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
318 ; SSE2-NEXT:    callq _roundevenf
319 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
320 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
321 ; SSE2-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload
322 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
323 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
324 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
325 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
326 ; SSE2-NEXT:    callq _roundevenf
327 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
328 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
329 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
330 ; SSE2-NEXT:    callq _roundevenf
331 ; SSE2-NEXT:    unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload
332 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
333 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
334 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
335 ; SSE2-NEXT:    callq _roundevenf
336 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
337 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
338 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
339 ; SSE2-NEXT:    callq _roundevenf
340 ; SSE2-NEXT:    movaps (%rsp), %xmm1 ## 16-byte Reload
341 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
342 ; SSE2-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload
343 ; SSE2-NEXT:    ## xmm1 = xmm1[0],mem[0]
344 ; SSE2-NEXT:    movaps %xmm1, (%rsp) ## 16-byte Spill
345 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
346 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
347 ; SSE2-NEXT:    callq _roundevenf
348 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
349 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
350 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
351 ; SSE2-NEXT:    callq _roundevenf
352 ; SSE2-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
353 ; SSE2-NEXT:    ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
354 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
355 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
356 ; SSE2-NEXT:    callq _roundevenf
357 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
358 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
359 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
360 ; SSE2-NEXT:    callq _roundevenf
361 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 ## 16-byte Reload
362 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1]
363 ; SSE2-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 ## 16-byte Folded Reload
364 ; SSE2-NEXT:    ## xmm3 = xmm3[0],mem[0]
365 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
366 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
367 ; SSE2-NEXT:    movaps (%rsp), %xmm2 ## 16-byte Reload
368 ; SSE2-NEXT:    addq $104, %rsp
369 ; SSE2-NEXT:    retq
394 ; SSE2-LABEL: roundeven_v8f64:
395 ; SSE2:       ## %bb.0:
396 ; SSE2-NEXT:    subq $88, %rsp
397 ; SSE2-NEXT:    .cfi_def_cfa_offset 96
398 ; SSE2-NEXT:    movaps %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
399 ; SSE2-NEXT:    movaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
400 ; SSE2-NEXT:    movaps %xmm1, (%rsp) ## 16-byte Spill
401 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
402 ; SSE2-NEXT:    callq _roundeven
403 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
404 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
405 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
406 ; SSE2-NEXT:    callq _roundeven
407 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
408 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
409 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
410 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
411 ; SSE2-NEXT:    callq _roundeven
412 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
413 ; SSE2-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
414 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
415 ; SSE2-NEXT:    callq _roundeven
416 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
417 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
418 ; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
419 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
420 ; SSE2-NEXT:    callq _roundeven
421 ; SSE2-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
422 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
423 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
424 ; SSE2-NEXT:    callq _roundeven
425 ; SSE2-NEXT:    movaps (%rsp), %xmm1 ## 16-byte Reload
426 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
427 ; SSE2-NEXT:    movaps %xmm1, (%rsp) ## 16-byte Spill
428 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
429 ; SSE2-NEXT:    callq _roundeven
430 ; SSE2-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
431 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
432 ; SSE2-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
433 ; SSE2-NEXT:    callq _roundeven
434 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 ## 16-byte Reload
435 ; SSE2-NEXT:    movlhps {{.*#+}} xmm3 = xmm3[0],xmm0[0]
436 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
437 ; SSE2-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
438 ; SSE2-NEXT:    movaps (%rsp), %xmm2 ## 16-byte Reload
439 ; SSE2-NEXT:    addq $88, %rsp
440 ; SSE2-NEXT:    retq