1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512er --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
3 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512er --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
5 define <16 x float> @test_rsqrt28_ps(<16 x float> %a0) {
6 ; CHECK-LABEL: test_rsqrt28_ps:
8 ; CHECK-NEXT: vrsqrt28ps {sae}, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x18,0xcc,0xc0]
9 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
10 %res = call <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float> %a0, <16 x float> zeroinitializer, i16 -1, i32 8)
14 define <16 x float> @test1_rsqrt28_ps(<16 x float> %a0, <16 x float> %a1) {
15 ; CHECK-LABEL: test1_rsqrt28_ps:
17 ; CHECK-NEXT: movw $6, %ax # encoding: [0x66,0xb8,0x06,0x00]
18 ; CHECK-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
19 ; CHECK-NEXT: vrsqrt28ps {sae}, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf2,0x7d,0x19,0xcc,0xc8]
20 ; CHECK-NEXT: vmovaps %zmm1, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
21 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
22 %res = call <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float> %a0, <16 x float> %a1, i16 6, i32 8)
26 define <16 x float> @test2_rsqrt28_ps(<16 x float> %a0) {
27 ; CHECK-LABEL: test2_rsqrt28_ps:
29 ; CHECK-NEXT: movw $6, %ax # encoding: [0x66,0xb8,0x06,0x00]
30 ; CHECK-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
31 ; CHECK-NEXT: vrsqrt28ps %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcc,0xc0]
32 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
33 %res = call <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float> %a0, <16 x float> undef, i16 6, i32 4)
37 define <16 x float> @test3_rsqrt28_ps(<16 x float> %a0) {
38 ; CHECK-LABEL: test3_rsqrt28_ps:
40 ; CHECK-NEXT: movw $6, %ax # encoding: [0x66,0xb8,0x06,0x00]
41 ; CHECK-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
42 ; CHECK-NEXT: vrsqrt28ps %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcc,0xc0]
43 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
44 %res = call <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float> %a0, <16 x float> zeroinitializer, i16 6, i32 4)
48 define <16 x float> @test4_rsqrt28_ps(<16 x float> %a0) {
49 ; CHECK-LABEL: test4_rsqrt28_ps:
51 ; CHECK-NEXT: movw $6, %ax # encoding: [0x66,0xb8,0x06,0x00]
52 ; CHECK-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
53 ; CHECK-NEXT: vrsqrt28ps {sae}, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x99,0xcc,0xc0]
54 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
55 %res = call <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float> %a0, <16 x float> undef, i16 6, i32 8)
59 declare <16 x float> @llvm.x86.avx512.rsqrt28.ps(<16 x float>, <16 x float>, i16, i32) nounwind readnone
61 define <16 x float> @test_rcp28_ps_512(<16 x float> %a0) {
62 ; CHECK-LABEL: test_rcp28_ps_512:
64 ; CHECK-NEXT: vrcp28ps {sae}, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x18,0xca,0xc0]
65 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
66 %res = call <16 x float> @llvm.x86.avx512.rcp28.ps(<16 x float> %a0, <16 x float> zeroinitializer, i16 -1, i32 8)
69 declare <16 x float> @llvm.x86.avx512.rcp28.ps(<16 x float>, <16 x float>, i16, i32) nounwind readnone
71 define <8 x double> @test_rcp28_pd_512(<8 x double> %a0) {
72 ; CHECK-LABEL: test_rcp28_pd_512:
74 ; CHECK-NEXT: vrcp28pd {sae}, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x18,0xca,0xc0]
75 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
76 %res = call <8 x double> @llvm.x86.avx512.rcp28.pd(<8 x double> %a0, <8 x double> zeroinitializer, i8 -1, i32 8)
79 declare <8 x double> @llvm.x86.avx512.rcp28.pd(<8 x double>, <8 x double>, i8, i32) nounwind readnone
81 define <16 x float> @test_exp2_ps_512(<16 x float> %a0) {
82 ; CHECK-LABEL: test_exp2_ps_512:
84 ; CHECK-NEXT: vexp2ps {sae}, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x18,0xc8,0xc0]
85 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
86 %res = call <16 x float> @llvm.x86.avx512.exp2.ps(<16 x float> %a0, <16 x float> zeroinitializer, i16 -1, i32 8)
89 declare <16 x float> @llvm.x86.avx512.exp2.ps(<16 x float>, <16 x float>, i16, i32) nounwind readnone
91 define <8 x double> @test_exp2_pd_512(<8 x double> %a0) {
92 ; CHECK-LABEL: test_exp2_pd_512:
94 ; CHECK-NEXT: vexp2pd {sae}, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x18,0xc8,0xc0]
95 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
96 %res = call <8 x double> @llvm.x86.avx512.exp2.pd(<8 x double> %a0, <8 x double> zeroinitializer, i8 -1, i32 8)
99 declare <8 x double> @llvm.x86.avx512.exp2.pd(<8 x double>, <8 x double>, i8, i32) nounwind readnone
101 define <4 x float> @test_rsqrt28_ss(<4 x float> %a0) {
102 ; CHECK-LABEL: test_rsqrt28_ss:
104 ; CHECK-NEXT: vrsqrt28ss {sae}, %xmm0, %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x18,0xcd,0xc0]
105 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
106 %res = call <4 x float> @llvm.x86.avx512.rsqrt28.ss(<4 x float> %a0, <4 x float> %a0, <4 x float> zeroinitializer, i8 -1, i32 8) ; <<4 x float>> [#uses=1]
109 declare <4 x float> @llvm.x86.avx512.rsqrt28.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) nounwind readnone
111 define <4 x float> @test_rcp28_ss(<4 x float> %a0) {
112 ; CHECK-LABEL: test_rcp28_ss:
114 ; CHECK-NEXT: vrcp28ss {sae}, %xmm0, %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x18,0xcb,0xc0]
115 ; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
116 %res = call <4 x float> @llvm.x86.avx512.rcp28.ss(<4 x float> %a0, <4 x float> %a0, <4 x float> zeroinitializer, i8 -1, i32 8) ; <<4 x float>> [#uses=1]
119 declare <4 x float> @llvm.x86.avx512.rcp28.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) nounwind readnone
121 define <4 x float> @test_rcp28_ss_load(<4 x float> %a0, ptr %a1ptr) {
122 ; X86-LABEL: test_rcp28_ss_load:
124 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
125 ; X86-NEXT: vrcp28ss (%eax), %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0xcb,0x00]
126 ; X86-NEXT: retl # encoding: [0xc3]
128 ; X64-LABEL: test_rcp28_ss_load:
130 ; X64-NEXT: vrcp28ss (%rdi), %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0xcb,0x07]
131 ; X64-NEXT: retq # encoding: [0xc3]
132 %a1 = load <4 x float>, ptr %a1ptr
133 %res = call <4 x float> @llvm.x86.avx512.rcp28.ss(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1, i32 4) ; <<4 x float>> [#uses=1]
137 define <4 x float> @test_rsqrt28_ss_load(<4 x float> %a0, ptr %a1ptr) {
138 ; X86-LABEL: test_rsqrt28_ss_load:
140 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
141 ; X86-NEXT: vrsqrt28ss (%eax), %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0xcd,0x00]
142 ; X86-NEXT: retl # encoding: [0xc3]
144 ; X64-LABEL: test_rsqrt28_ss_load:
146 ; X64-NEXT: vrsqrt28ss (%rdi), %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0xcd,0x07]
147 ; X64-NEXT: retq # encoding: [0xc3]
148 %a1 = load <4 x float>, ptr %a1ptr
149 %res = call <4 x float> @llvm.x86.avx512.rsqrt28.ss(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1, i32 4) ; <<4 x float>> [#uses=1]
153 define <4 x float> @test_rsqrt28_ss_maskz(<4 x float> %a0, i8 %mask) {
154 ; X86-LABEL: test_rsqrt28_ss_maskz:
156 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
157 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
158 ; X86-NEXT: vrsqrt28ss {sae}, %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x99,0xcd,0xc0]
159 ; X86-NEXT: retl # encoding: [0xc3]
161 ; X64-LABEL: test_rsqrt28_ss_maskz:
163 ; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
164 ; X64-NEXT: vrsqrt28ss {sae}, %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x99,0xcd,0xc0]
165 ; X64-NEXT: retq # encoding: [0xc3]
166 %res = call <4 x float> @llvm.x86.avx512.rsqrt28.ss(<4 x float> %a0, <4 x float> %a0, <4 x float> zeroinitializer, i8 %mask, i32 8) ;
170 define <4 x float> @test_rsqrt28_ss_mask(<4 x float> %a0, <4 x float> %b0, <4 x float> %c0, i8 %mask) {
171 ; X86-LABEL: test_rsqrt28_ss_mask:
173 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
174 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
175 ; X86-NEXT: vrsqrt28ss {sae}, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x19,0xcd,0xd1]
176 ; X86-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
177 ; X86-NEXT: retl # encoding: [0xc3]
179 ; X64-LABEL: test_rsqrt28_ss_mask:
181 ; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
182 ; X64-NEXT: vrsqrt28ss {sae}, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x19,0xcd,0xd1]
183 ; X64-NEXT: vmovaps %xmm2, %xmm0 # encoding: [0xc5,0xf8,0x28,0xc2]
184 ; X64-NEXT: retq # encoding: [0xc3]
185 %res = call <4 x float> @llvm.x86.avx512.rsqrt28.ss(<4 x float> %a0, <4 x float> %b0, <4 x float> %c0, i8 %mask, i32 8) ;
189 define <2 x double> @test_rcp28_sd_mask_load(<2 x double> %a0, ptr %a1ptr, <2 x double> %a2, i8 %mask) {
190 ; X86-LABEL: test_rcp28_sd_mask_load:
192 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x08]
193 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
194 ; X86-NEXT: vrcp28sd %xmm0, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xcb,0xc8]
195 ; X86-NEXT: vmovapd %xmm1, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc1]
196 ; X86-NEXT: retl # encoding: [0xc3]
198 ; X64-LABEL: test_rcp28_sd_mask_load:
200 ; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
201 ; X64-NEXT: vrcp28sd %xmm0, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0xcb,0xc8]
202 ; X64-NEXT: vmovapd %xmm1, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc1]
203 ; X64-NEXT: retq # encoding: [0xc3]
204 %a1 = load <2 x double>, ptr %a1ptr
205 %res = call <2 x double> @llvm.x86.avx512.rcp28.sd(<2 x double> %a0, <2 x double> %a0, <2 x double> %a2, i8 %mask, i32 4) ;
206 ret <2 x double> %res
208 declare <2 x double> @llvm.x86.avx512.rcp28.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) nounwind readnone
210 define <2 x double> @test_rsqrt28_sd_maskz_load(<2 x double> %a0, ptr %a1ptr, i8 %mask) {
211 ; X86-LABEL: test_rsqrt28_sd_maskz_load:
213 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x08]
214 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
215 ; X86-NEXT: vrsqrt28sd %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0xc0]
216 ; X86-NEXT: retl # encoding: [0xc3]
218 ; X64-LABEL: test_rsqrt28_sd_maskz_load:
220 ; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
221 ; X64-NEXT: vrsqrt28sd %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0xc0]
222 ; X64-NEXT: retq # encoding: [0xc3]
223 %a1 = load <2 x double>, ptr %a1ptr
224 %res = call <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double> %a0, <2 x double> %a0, <2 x double> zeroinitializer, i8 %mask, i32 4) ;
225 ret <2 x double> %res
228 define <2 x double> @test_rsqrt28_sd_maskz(<2 x double> %a0, i8 %mask) {
229 ; X86-LABEL: test_rsqrt28_sd_maskz:
231 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
232 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
233 ; X86-NEXT: vrsqrt28sd {sae}, %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x99,0xcd,0xc0]
234 ; X86-NEXT: retl # encoding: [0xc3]
236 ; X64-LABEL: test_rsqrt28_sd_maskz:
238 ; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
239 ; X64-NEXT: vrsqrt28sd {sae}, %xmm0, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x99,0xcd,0xc0]
240 ; X64-NEXT: retq # encoding: [0xc3]
241 %res = call <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double> %a0, <2 x double> %a0, <2 x double> zeroinitializer, i8 %mask, i32 8) ;
242 ret <2 x double> %res
245 define <2 x double> @test_rsqrt28_sd_mask(<2 x double> %a0, <2 x double> %b0, <2 x double> %c0, i8 %mask) {
246 ; X86-LABEL: test_rsqrt28_sd_mask:
248 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
249 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
250 ; X86-NEXT: vrsqrt28sd {sae}, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x19,0xcd,0xd1]
251 ; X86-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
252 ; X86-NEXT: retl # encoding: [0xc3]
254 ; X64-LABEL: test_rsqrt28_sd_mask:
256 ; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
257 ; X64-NEXT: vrsqrt28sd {sae}, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x19,0xcd,0xd1]
258 ; X64-NEXT: vmovapd %xmm2, %xmm0 # encoding: [0xc5,0xf9,0x28,0xc2]
259 ; X64-NEXT: retq # encoding: [0xc3]
260 %res = call <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double> %a0, <2 x double> %b0, <2 x double> %c0, i8 %mask, i32 8) ;
261 ret <2 x double> %res
264 declare <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) nounwind readnone
266 define <2 x double> @test_rsqrt28_sd_maskz_mem(<2 x double> %a0, ptr %ptr, i8 %mask) {
267 ; X86-LABEL: test_rsqrt28_sd_maskz_mem:
269 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x08]
270 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
271 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
272 ; X86-NEXT: vrsqrt28sd (%eax), %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0x00]
273 ; X86-NEXT: retl # encoding: [0xc3]
275 ; X64-LABEL: test_rsqrt28_sd_maskz_mem:
277 ; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
278 ; X64-NEXT: vrsqrt28sd (%rdi), %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0x07]
279 ; X64-NEXT: retq # encoding: [0xc3]
280 %mem = load double , ptr %ptr, align 8
281 %mem_v = insertelement <2 x double> undef, double %mem, i32 0
282 %res = call <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double> %a0, <2 x double> %mem_v, <2 x double> zeroinitializer, i8 %mask, i32 4) ;
283 ret <2 x double> %res
286 define <2 x double> @test_rsqrt28_sd_maskz_mem_offset(<2 x double> %a0, ptr %ptr, i8 %mask) {
287 ; X86-LABEL: test_rsqrt28_sd_maskz_mem_offset:
289 ; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x08]
290 ; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
291 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
292 ; X86-NEXT: vrsqrt28sd 144(%eax), %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0x40,0x12]
293 ; X86-NEXT: retl # encoding: [0xc3]
295 ; X64-LABEL: test_rsqrt28_sd_maskz_mem_offset:
297 ; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
298 ; X64-NEXT: vrsqrt28sd 144(%rdi), %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0xcd,0x47,0x12]
299 ; X64-NEXT: retq # encoding: [0xc3]
300 %ptr1 = getelementptr double, ptr %ptr, i32 18
301 %mem = load double , ptr %ptr1, align 8
302 %mem_v = insertelement <2 x double> undef, double %mem, i32 0
303 %res = call <2 x double> @llvm.x86.avx512.rsqrt28.sd(<2 x double> %a0, <2 x double> %mem_v, <2 x double> zeroinitializer, i8 %mask, i32 4) ;
304 ret <2 x double> %res