@@ -16,6 +16,7 @@ target triple = "x86_64-apple-macosx10.8.0"
16
16
define i32 @ceil (i32 %arg ) {
17
17
; SSE2-LABEL: 'ceil'
18
18
; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %F32 = call float @llvm.ceil.f32(float undef)
19
+ ; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2F32 = call <2 x float> @llvm.ceil.v2f32(<2 x float> undef)
19
20
; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V4F32 = call <4 x float> @llvm.ceil.v4f32(<4 x float> undef)
20
21
; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V8F32 = call <8 x float> @llvm.ceil.v8f32(<8 x float> undef)
21
22
; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef)
@@ -27,6 +28,7 @@ define i32 @ceil(i32 %arg) {
27
28
;
28
29
; SSE42-LABEL: 'ceil'
29
30
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.ceil.f32(float undef)
31
+ ; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.ceil.v2f32(<2 x float> undef)
30
32
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.ceil.v4f32(<4 x float> undef)
31
33
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = call <8 x float> @llvm.ceil.v8f32(<8 x float> undef)
32
34
; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef)
@@ -38,6 +40,7 @@ define i32 @ceil(i32 %arg) {
38
40
;
39
41
; AVX-LABEL: 'ceil'
40
42
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.ceil.f32(float undef)
43
+ ; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.ceil.v2f32(<2 x float> undef)
41
44
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.ceil.v4f32(<4 x float> undef)
42
45
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.ceil.v8f32(<8 x float> undef)
43
46
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef)
@@ -49,6 +52,7 @@ define i32 @ceil(i32 %arg) {
49
52
;
50
53
; AVX512-LABEL: 'ceil'
51
54
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.ceil.f32(float undef)
55
+ ; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.ceil.v2f32(<2 x float> undef)
52
56
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.ceil.v4f32(<4 x float> undef)
53
57
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.ceil.v8f32(<8 x float> undef)
54
58
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef)
@@ -59,6 +63,7 @@ define i32 @ceil(i32 %arg) {
59
63
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
60
64
;
61
65
%F32 = call float @llvm.ceil.f32 (float undef )
66
+ %V2F32 = call <2 x float > @llvm.ceil.v2f32 (<2 x float > undef )
62
67
%V4F32 = call <4 x float > @llvm.ceil.v4f32 (<4 x float > undef )
63
68
%V8F32 = call <8 x float > @llvm.ceil.v8f32 (<8 x float > undef )
64
69
%V16F32 = call <16 x float > @llvm.ceil.v16f32 (<16 x float > undef )
@@ -74,6 +79,7 @@ define i32 @ceil(i32 %arg) {
74
79
define i32 @floor (i32 %arg ) {
75
80
; SSE2-LABEL: 'floor'
76
81
; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %F32 = call float @llvm.floor.f32(float undef)
82
+ ; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2F32 = call <2 x float> @llvm.floor.v2f32(<2 x float> undef)
77
83
; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V4F32 = call <4 x float> @llvm.floor.v4f32(<4 x float> undef)
78
84
; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V8F32 = call <8 x float> @llvm.floor.v8f32(<8 x float> undef)
79
85
; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef)
@@ -85,6 +91,7 @@ define i32 @floor(i32 %arg) {
85
91
;
86
92
; SSE42-LABEL: 'floor'
87
93
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.floor.f32(float undef)
94
+ ; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.floor.v2f32(<2 x float> undef)
88
95
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.floor.v4f32(<4 x float> undef)
89
96
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = call <8 x float> @llvm.floor.v8f32(<8 x float> undef)
90
97
; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef)
@@ -96,6 +103,7 @@ define i32 @floor(i32 %arg) {
96
103
;
97
104
; AVX-LABEL: 'floor'
98
105
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.floor.f32(float undef)
106
+ ; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.floor.v2f32(<2 x float> undef)
99
107
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.floor.v4f32(<4 x float> undef)
100
108
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.floor.v8f32(<8 x float> undef)
101
109
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef)
@@ -107,6 +115,7 @@ define i32 @floor(i32 %arg) {
107
115
;
108
116
; AVX512-LABEL: 'floor'
109
117
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.floor.f32(float undef)
118
+ ; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.floor.v2f32(<2 x float> undef)
110
119
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.floor.v4f32(<4 x float> undef)
111
120
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.floor.v8f32(<8 x float> undef)
112
121
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef)
@@ -117,6 +126,7 @@ define i32 @floor(i32 %arg) {
117
126
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
118
127
;
119
128
%F32 = call float @llvm.floor.f32 (float undef )
129
+ %V2F32 = call <2 x float > @llvm.floor.v2f32 (<2 x float > undef )
120
130
%V4F32 = call <4 x float > @llvm.floor.v4f32 (<4 x float > undef )
121
131
%V8F32 = call <8 x float > @llvm.floor.v8f32 (<8 x float > undef )
122
132
%V16F32 = call <16 x float > @llvm.floor.v16f32 (<16 x float > undef )
@@ -132,6 +142,7 @@ define i32 @floor(i32 %arg) {
132
142
define i32 @nearbyint (i32 %arg ) {
133
143
; SSE2-LABEL: 'nearbyint'
134
144
; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %F32 = call float @llvm.nearbyint.f32(float undef)
145
+ ; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2F32 = call <2 x float> @llvm.nearbyint.v2f32(<2 x float> undef)
135
146
; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V4F32 = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> undef)
136
147
; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V8F32 = call <8 x float> @llvm.nearbyint.v8f32(<8 x float> undef)
137
148
; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.nearbyint.v16f32(<16 x float> undef)
@@ -143,6 +154,7 @@ define i32 @nearbyint(i32 %arg) {
143
154
;
144
155
; SSE42-LABEL: 'nearbyint'
145
156
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.nearbyint.f32(float undef)
157
+ ; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.nearbyint.v2f32(<2 x float> undef)
146
158
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> undef)
147
159
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = call <8 x float> @llvm.nearbyint.v8f32(<8 x float> undef)
148
160
; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.nearbyint.v16f32(<16 x float> undef)
@@ -154,6 +166,7 @@ define i32 @nearbyint(i32 %arg) {
154
166
;
155
167
; AVX-LABEL: 'nearbyint'
156
168
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.nearbyint.f32(float undef)
169
+ ; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.nearbyint.v2f32(<2 x float> undef)
157
170
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> undef)
158
171
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.nearbyint.v8f32(<8 x float> undef)
159
172
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.nearbyint.v16f32(<16 x float> undef)
@@ -165,6 +178,7 @@ define i32 @nearbyint(i32 %arg) {
165
178
;
166
179
; AVX512-LABEL: 'nearbyint'
167
180
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.nearbyint.f32(float undef)
181
+ ; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.nearbyint.v2f32(<2 x float> undef)
168
182
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> undef)
169
183
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.nearbyint.v8f32(<8 x float> undef)
170
184
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.nearbyint.v16f32(<16 x float> undef)
@@ -175,6 +189,7 @@ define i32 @nearbyint(i32 %arg) {
175
189
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
176
190
;
177
191
%F32 = call float @llvm.nearbyint.f32 (float undef )
192
+ %V2F32 = call <2 x float > @llvm.nearbyint.v2f32 (<2 x float > undef )
178
193
%V4F32 = call <4 x float > @llvm.nearbyint.v4f32 (<4 x float > undef )
179
194
%V8F32 = call <8 x float > @llvm.nearbyint.v8f32 (<8 x float > undef )
180
195
%V16F32 = call <16 x float > @llvm.nearbyint.v16f32 (<16 x float > undef )
@@ -190,6 +205,7 @@ define i32 @nearbyint(i32 %arg) {
190
205
define i32 @rint (i32 %arg ) {
191
206
; SSE2-LABEL: 'rint'
192
207
; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %F32 = call float @llvm.rint.f32(float undef)
208
+ ; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2F32 = call <2 x float> @llvm.rint.v2f32(<2 x float> undef)
193
209
; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V4F32 = call <4 x float> @llvm.rint.v4f32(<4 x float> undef)
194
210
; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V8F32 = call <8 x float> @llvm.rint.v8f32(<8 x float> undef)
195
211
; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.rint.v16f32(<16 x float> undef)
@@ -201,6 +217,7 @@ define i32 @rint(i32 %arg) {
201
217
;
202
218
; SSE42-LABEL: 'rint'
203
219
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.rint.f32(float undef)
220
+ ; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.rint.v2f32(<2 x float> undef)
204
221
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.rint.v4f32(<4 x float> undef)
205
222
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = call <8 x float> @llvm.rint.v8f32(<8 x float> undef)
206
223
; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.rint.v16f32(<16 x float> undef)
@@ -212,6 +229,7 @@ define i32 @rint(i32 %arg) {
212
229
;
213
230
; AVX-LABEL: 'rint'
214
231
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.rint.f32(float undef)
232
+ ; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.rint.v2f32(<2 x float> undef)
215
233
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.rint.v4f32(<4 x float> undef)
216
234
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.rint.v8f32(<8 x float> undef)
217
235
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.rint.v16f32(<16 x float> undef)
@@ -223,6 +241,7 @@ define i32 @rint(i32 %arg) {
223
241
;
224
242
; AVX512-LABEL: 'rint'
225
243
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.rint.f32(float undef)
244
+ ; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.rint.v2f32(<2 x float> undef)
226
245
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.rint.v4f32(<4 x float> undef)
227
246
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.rint.v8f32(<8 x float> undef)
228
247
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.rint.v16f32(<16 x float> undef)
@@ -233,6 +252,7 @@ define i32 @rint(i32 %arg) {
233
252
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
234
253
;
235
254
%F32 = call float @llvm.rint.f32 (float undef )
255
+ %V2F32 = call <2 x float > @llvm.rint.v2f32 (<2 x float > undef )
236
256
%V4F32 = call <4 x float > @llvm.rint.v4f32 (<4 x float > undef )
237
257
%V8F32 = call <8 x float > @llvm.rint.v8f32 (<8 x float > undef )
238
258
%V16F32 = call <16 x float > @llvm.rint.v16f32 (<16 x float > undef )
@@ -248,6 +268,7 @@ define i32 @rint(i32 %arg) {
248
268
define i32 @trunc (i32 %arg ) {
249
269
; SSE2-LABEL: 'trunc'
250
270
; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %F32 = call float @llvm.trunc.f32(float undef)
271
+ ; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2F32 = call <2 x float> @llvm.trunc.v2f32(<2 x float> undef)
251
272
; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V4F32 = call <4 x float> @llvm.trunc.v4f32(<4 x float> undef)
252
273
; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V8F32 = call <8 x float> @llvm.trunc.v8f32(<8 x float> undef)
253
274
; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.trunc.v16f32(<16 x float> undef)
@@ -259,6 +280,7 @@ define i32 @trunc(i32 %arg) {
259
280
;
260
281
; SSE42-LABEL: 'trunc'
261
282
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.trunc.f32(float undef)
283
+ ; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.trunc.v2f32(<2 x float> undef)
262
284
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.trunc.v4f32(<4 x float> undef)
263
285
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = call <8 x float> @llvm.trunc.v8f32(<8 x float> undef)
264
286
; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.trunc.v16f32(<16 x float> undef)
@@ -270,6 +292,7 @@ define i32 @trunc(i32 %arg) {
270
292
;
271
293
; AVX-LABEL: 'trunc'
272
294
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.trunc.f32(float undef)
295
+ ; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.trunc.v2f32(<2 x float> undef)
273
296
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.trunc.v4f32(<4 x float> undef)
274
297
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.trunc.v8f32(<8 x float> undef)
275
298
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.trunc.v16f32(<16 x float> undef)
@@ -281,6 +304,7 @@ define i32 @trunc(i32 %arg) {
281
304
;
282
305
; AVX512-LABEL: 'trunc'
283
306
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call float @llvm.trunc.f32(float undef)
307
+ ; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = call <2 x float> @llvm.trunc.v2f32(<2 x float> undef)
284
308
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x float> @llvm.trunc.v4f32(<4 x float> undef)
285
309
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x float> @llvm.trunc.v8f32(<8 x float> undef)
286
310
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.trunc.v16f32(<16 x float> undef)
@@ -291,6 +315,7 @@ define i32 @trunc(i32 %arg) {
291
315
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
292
316
;
293
317
%F32 = call float @llvm.trunc.f32 (float undef )
318
+ %V2F32 = call <2 x float > @llvm.trunc.v2f32 (<2 x float > undef )
294
319
%V4F32 = call <4 x float > @llvm.trunc.v4f32 (<4 x float > undef )
295
320
%V8F32 = call <8 x float > @llvm.trunc.v8f32 (<8 x float > undef )
296
321
%V16F32 = call <16 x float > @llvm.trunc.v16f32 (<16 x float > undef )
@@ -304,6 +329,7 @@ define i32 @trunc(i32 %arg) {
304
329
}
305
330
306
331
declare float @llvm.ceil.f32 (float )
332
+ declare <2 x float > @llvm.ceil.v2f32 (<2 x float >)
307
333
declare <4 x float > @llvm.ceil.v4f32 (<4 x float >)
308
334
declare <8 x float > @llvm.ceil.v8f32 (<8 x float >)
309
335
declare <16 x float > @llvm.ceil.v16f32 (<16 x float >)
@@ -314,6 +340,7 @@ declare <4 x double> @llvm.ceil.v4f64(<4 x double>)
314
340
declare <8 x double > @llvm.ceil.v8f64 (<8 x double >)
315
341
316
342
declare float @llvm.floor.f32 (float )
343
+ declare <2 x float > @llvm.floor.v2f32 (<2 x float >)
317
344
declare <4 x float > @llvm.floor.v4f32 (<4 x float >)
318
345
declare <8 x float > @llvm.floor.v8f32 (<8 x float >)
319
346
declare <16 x float > @llvm.floor.v16f32 (<16 x float >)
@@ -324,6 +351,7 @@ declare <4 x double> @llvm.floor.v4f64(<4 x double>)
324
351
declare <8 x double > @llvm.floor.v8f64 (<8 x double >)
325
352
326
353
declare float @llvm.nearbyint.f32 (float )
354
+ declare <2 x float > @llvm.nearbyint.v2f32 (<2 x float >)
327
355
declare <4 x float > @llvm.nearbyint.v4f32 (<4 x float >)
328
356
declare <8 x float > @llvm.nearbyint.v8f32 (<8 x float >)
329
357
declare <16 x float > @llvm.nearbyint.v16f32 (<16 x float >)
@@ -334,6 +362,7 @@ declare <4 x double> @llvm.nearbyint.v4f64(<4 x double>)
334
362
declare <8 x double > @llvm.nearbyint.v8f64 (<8 x double >)
335
363
336
364
declare float @llvm.rint.f32 (float )
365
+ declare <2 x float > @llvm.rint.v2f32 (<2 x float >)
337
366
declare <4 x float > @llvm.rint.v4f32 (<4 x float >)
338
367
declare <8 x float > @llvm.rint.v8f32 (<8 x float >)
339
368
declare <16 x float > @llvm.rint.v16f32 (<16 x float >)
@@ -344,6 +373,7 @@ declare <4 x double> @llvm.rint.v4f64(<4 x double>)
344
373
declare <8 x double > @llvm.rint.v8f64 (<8 x double >)
345
374
346
375
declare float @llvm.trunc.f32 (float )
376
+ declare <2 x float > @llvm.trunc.v2f32 (<2 x float >)
347
377
declare <4 x float > @llvm.trunc.v4f32 (<4 x float >)
348
378
declare <8 x float > @llvm.trunc.v8f32 (<8 x float >)
349
379
declare <16 x float > @llvm.trunc.v16f32 (<16 x float >)
0 commit comments