[X86] Adding avx512_vpopcntdq feature set and its intrinsics

[oren_ben_simhon]

Summary

AVX512_VPOPCNTDQ is a new feature set that Intel published here.

The patch represents the Clang side of the addition of six intrinsics for two new machine instructions (vpopcntd and vpopcntq).
It also includes the addition of the new feature set.

[RKSimon]

include/clang/Basic/BuiltinsX86.def

1105

Would we be better off directly emitting Intrinsic::ctpop inside CGBuiltin.cpp?

AFAICT these builtins have no additional abilities compared to ctpop (other than masking which we can already handle).

[oren_ben_simhon]

Removed files properties

[craig.topper]

Looks like gcc already implements this https://www.mail-archive.com/[e-mail]/msg156421.html

Looks like their command line switch is -mavx512vpopcntdq and their preprocessor define is AVX512VPOPCNTDQ without an underscore. We should match for compatibility.

[oren_ben_simhon]

Implemented comments posted until 05/15 (Thanks Craig and Simon)

[craig.topper]

lib/CodeGen/CGBuiltin.cpp

7500

I'm not sure what EmitScalarExpr does, but I got think its not right for a vector argument.

7501

I think you can use Ops[0]->getType() and avoid the call to ConvertType. See the code for BI__builtin_ia32_vplzcntq_512_mask.

[oren_ben_simhon]

Implemented comments posted until 05/16 (Thanks Craig)

lib/CodeGen/CGBuiltin.cpp

7500

Vector type is considered as scalar (single value) type from the emitter point of view.
See also: http://llvm.org/docs/LangRef.html#single-value-types

[craig.topper]

lib/CodeGen/CGBuiltin.cpp

7500

Why can't we just use Ops[0] in place of X?

[oren_ben_simhon]

Implemented comments posted until 05/20 (Thanks Craig)

[craig.topper]

lib/CodeGen/CGBuiltin.cpp

7500

Why did the call to ConvertType come back? This code can just be

llvm::Function *F = CGM.getIntrinsic(Intrinsic::ctpop, Ops[0]->getType());
return Builder.CreateCall(F, Ops);

[oren_ben_simhon]

lib/CodeGen/CGBuiltin.cpp

7500

After checking it again, I reverted that change because AFAIK Ops[0] is the first argument and not the return type of the statement.

7501

After checking it again, I am going to revert this change because AFAIK Ops[0] is the first argument and the return type.

[oren_ben_simhon]

lib/CodeGen/CGBuiltin.cpp

7500

After checking it again, I reverted that change because AFAIK Ops[0] is the first argument and not the return type of the statement.

[craig.topper]

LGTM

lib/CodeGen/CGBuiltin.cpp

7500

That's true, but for this intrinsic they are required to be the same type. But its not a big deal.

include/clang/Basic/BuiltinsX86.def

 TARGET_BUILTIN(__builtin_ia32_vpconflictdi_256_mask, "V4LLiV4LLiV4LLiUc","","avx512cd,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_vpconflictsi_128_mask, "V4iV4iV4iUc","","avx512cd,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_vpconflictsi_256_mask, "V8iV8iV8iUc","","avx512cd,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_vpconflictdi_512_mask, "V8LLiV8LLiV8LLiUc", "", "avx512cd")
 TARGET_BUILTIN(__builtin_ia32_vpconflictsi_512_mask, "V16iV16iV16iUs", "", "avx512cd")
 TARGET_BUILTIN(__builtin_ia32_vplzcntd_512_mask, "V16iV16iV16iUs", "", "avx512cd")
 TARGET_BUILTIN(__builtin_ia32_vplzcntq_512_mask, "V8LLiV8LLiV8LLiUc", "", "avx512cd")
 
+TARGET_BUILTIN(__builtin_ia32_vpopcntd_512_mask, "V16iV16iV16iUs", "", "avx512_vpopcntdq")
+TARGET_BUILTIN(__builtin_ia32_vpopcntq_512_mask, "V8LLiV8LLiV8LLiUc", "", "avx512_vpopcntdq")
+
 TARGET_BUILTIN(__builtin_ia32_vpermi2varhi128_mask, "V8sV8sV8sV8sUc", "", "avx512vl,avx512bw")
 TARGET_BUILTIN(__builtin_ia32_vpermi2varhi256_mask, "V16sV16sV16sV16sUs", "", "avx512vl,avx512bw")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varhi128_mask, "V8sV8sV8sV8sUc", "", "avx512vl,avx512bw")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varhi128_maskz, "V8sV8sV8sV8sUc", "", "avx512vl,avx512bw")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varhi256_mask, "V16sV16sV16sV16sUs", "", "avx512vl,avx512bw")
 TARGET_BUILTIN(__builtin_ia32_vpermt2varhi256_maskz, "V16sV16sV16sV16sUs", "", "avx512vl,avx512bw")
 
 TARGET_BUILTIN(__builtin_ia32_pmulhrsw512_mask, "V32sV32sV32sV32sUi", "", "avx512bw")

include/clang/Driver/Options.td

 def mno_sse4 : Flag<["-"], "mno-sse4">, Alias<mno_sse4_1>;
 def mno_sse : Flag<["-"], "mno-sse">, Group<m_x86_Features_Group>;
 def mno_ssse3 : Flag<["-"], "mno-ssse3">, Group<m_x86_Features_Group>;
 def mno_aes : Flag<["-"], "mno-aes">, Group<m_x86_Features_Group>;
 def mno_avx : Flag<["-"], "mno-avx">, Group<m_x86_Features_Group>;
 def mno_avx2 : Flag<["-"], "mno-avx2">, Group<m_x86_Features_Group>;
 def mno_avx512f : Flag<["-"], "mno-avx512f">, Group<m_x86_Features_Group>;
 def mno_avx512cd : Flag<["-"], "mno-avx512cd">, Group<m_x86_Features_Group>;
+def mno_avx512_vpopcntdq : Flag<["-"], "mno-avx512_vpopcntdq">, Group<m_x86_Features_Group>;
 def mno_avx512er : Flag<["-"], "mno-avx512er">, Group<m_x86_Features_Group>;
 def mno_avx512pf : Flag<["-"], "mno-avx512pf">, Group<m_x86_Features_Group>;
 def mno_avx512dq : Flag<["-"], "mno-avx512dq">, Group<m_x86_Features_Group>;
 def mno_avx512bw : Flag<["-"], "mno-avx512bw">, Group<m_x86_Features_Group>;
 def mno_avx512vl : Flag<["-"], "mno-avx512vl">, Group<m_x86_Features_Group>;
 def mno_avx512vbmi : Flag<["-"], "mno-avx512vbmi">, Group<m_x86_Features_Group>;
 def mno_avx512ifma : Flag<["-"], "mno-avx512ifma">, Group<m_x86_Features_Group>;
 def mno_pclmul : Flag<["-"], "mno-pclmul">, Group<m_x86_Features_Group>;

 def msse4 : Flag<["-"], "msse4">, Alias<msse4_2>;
 def msse : Flag<["-"], "msse">, Group<m_x86_Features_Group>;
 def mssse3 : Flag<["-"], "mssse3">, Group<m_x86_Features_Group>;
 def maes : Flag<["-"], "maes">, Group<m_x86_Features_Group>;
 def mavx : Flag<["-"], "mavx">, Group<m_x86_Features_Group>;
 def mavx2 : Flag<["-"], "mavx2">, Group<m_x86_Features_Group>;
 def mavx512f : Flag<["-"], "mavx512f">, Group<m_x86_Features_Group>;
 def mavx512cd : Flag<["-"], "mavx512cd">, Group<m_x86_Features_Group>;
+def mavx512_vpopcntdq : Flag<["-"], "mavx512_vpopcntdq">, Group<m_x86_Features_Group>;
 def mavx512er : Flag<["-"], "mavx512er">, Group<m_x86_Features_Group>;
 def mavx512pf : Flag<["-"], "mavx512pf">, Group<m_x86_Features_Group>;
 def mavx512dq : Flag<["-"], "mavx512dq">, Group<m_x86_Features_Group>;
 def mavx512bw : Flag<["-"], "mavx512bw">, Group<m_x86_Features_Group>;
 def mavx512vl : Flag<["-"], "mavx512vl">, Group<m_x86_Features_Group>;
 def mavx512vbmi : Flag<["-"], "mavx512vbmi">, Group<m_x86_Features_Group>;
 def mavx512ifma : Flag<["-"], "mavx512ifma">, Group<m_x86_Features_Group>;
 def mpclmul : Flag<["-"], "mpclmul">, Group<m_x86_Features_Group>;

lib/Basic/Targets.cpp

class X86TargetInfo : public TargetInfo {
   bool HasRTM = false;
   bool HasPRFCHW = false;
   bool HasRDSEED = false;
   bool HasADX = false;
   bool HasTBM = false;
   bool HasFMA = false;
   bool HasF16C = false;
   bool HasAVX512CD = false;
+  bool HasAVX512_VPOPCNTDQ = false;
   bool HasAVX512ER = false;
   bool HasAVX512PF = false;
   bool HasAVX512DQ = false;
   bool HasAVX512BW = false;
   bool HasAVX512VL = false;
   bool HasAVX512VBMI = false;
   bool HasAVX512IFMA = false;
   bool HasSHA = false;
void X86TargetInfo::setSSELevel(llvm::StringMap<bool> &Features,
   case AVX:
     Features["fma"] = Features["avx"] = Features["f16c"] = Features["xsave"] =
       Features["xsaveopt"] = false;
     setXOPLevel(Features, FMA4, false);
   case AVX2:
     Features["avx2"] = false;
   case AVX512F:
     Features["avx512f"] = Features["avx512cd"] = Features["avx512er"] =
         Features["avx512pf"] = Features["avx512dq"] = Features["avx512bw"] =
             Features["avx512vl"] = Features["avx512vbmi"] =
-      Features["avx512ifma"] = false;
+                Features["avx512ifma"] = Features["avx512_vpopcntdq"] = false;
   }
 }
 
 void X86TargetInfo::setMMXLevel(llvm::StringMap<bool> &Features,
                                 MMX3DNowEnum Level, bool Enabled) {
   if (Enabled) {
     switch (Level) {
     case AMD3DNowAthlon:
void X86TargetInfo::setFeatureEnabledImpl(llvm::StringMap<bool> &Features,
   } else if (Name == "avx") {
     setSSELevel(Features, AVX, Enabled);
   } else if (Name == "avx2") {
     setSSELevel(Features, AVX2, Enabled);
   } else if (Name == "avx512f") {
     setSSELevel(Features, AVX512F, Enabled);
   } else if (Name == "avx512cd" || Name == "avx512er" || Name == "avx512pf" ||
              Name == "avx512dq" || Name == "avx512bw" || Name == "avx512vl" ||
-             Name == "avx512vbmi" || Name == "avx512ifma") {
+             Name == "avx512vbmi" || Name == "avx512ifma" ||
+             Name == "avx512_vpopcntdq") {
     if (Enabled)
       setSSELevel(Features, AVX512F, Enabled);
     // Enable BWI instruction if VBMI is being enabled.
     if (Name == "avx512vbmi" && Enabled)
       Features["avx512bw"] = true;
     // Also disable VBMI if BWI is being disabled.
     if (Name == "avx512bw" && !Enabled)
       Features["avx512vbmi"] = false;
for (const auto &Feature : Features) {
     } else if (Feature == "+tbm") {
       HasTBM = true;
     } else if (Feature == "+fma") {
       HasFMA = true;
     } else if (Feature == "+f16c") {
       HasF16C = true;
     } else if (Feature == "+avx512cd") {
       HasAVX512CD = true;
+    } else if (Feature == "+avx512_vpopcntdq") {
+      HasAVX512_VPOPCNTDQ = true;
     } else if (Feature == "+avx512er") {
       HasAVX512ER = true;
     } else if (Feature == "+avx512pf") {
       HasAVX512PF = true;
     } else if (Feature == "+avx512dq") {
       HasAVX512DQ = true;
     } else if (Feature == "+avx512bw") {
       HasAVX512BW = true;
void X86TargetInfo::getTargetDefines(const LangOptions &Opts,
   if (HasFMA)
     Builder.defineMacro("__FMA__");
 
   if (HasF16C)
     Builder.defineMacro("__F16C__");
 
   if (HasAVX512CD)
     Builder.defineMacro("__AVX512CD__");
+  if (HasAVX512_VPOPCNTDQ)
+    Builder.defineMacro("__AVX512_VPOPCNTDQ__");
   if (HasAVX512ER)
     Builder.defineMacro("__AVX512ER__");
   if (HasAVX512PF)
     Builder.defineMacro("__AVX512PF__");
   if (HasAVX512DQ)
     Builder.defineMacro("__AVX512DQ__");
   if (HasAVX512BW)
     Builder.defineMacro("__AVX512BW__");

 
 bool X86TargetInfo::hasFeature(StringRef Feature) const {
   return llvm::StringSwitch<bool>(Feature)
       .Case("aes", HasAES)
       .Case("avx", SSELevel >= AVX)
       .Case("avx2", SSELevel >= AVX2)
       .Case("avx512f", SSELevel >= AVX512F)
       .Case("avx512cd", HasAVX512CD)
+      .Case("avx512_vpopcntdq", HasAVX512_VPOPCNTDQ)
       .Case("avx512er", HasAVX512ER)
       .Case("avx512pf", HasAVX512PF)
       .Case("avx512dq", HasAVX512DQ)
       .Case("avx512bw", HasAVX512BW)
       .Case("avx512vl", HasAVX512VL)
       .Case("avx512vbmi", HasAVX512VBMI)
       .Case("avx512ifma", HasAVX512IFMA)
       .Case("bmi", HasBMI)
return llvm::StringSwitch<bool>(FeatureStr)
       .Case("bmi", true)
       .Case("bmi2", true)
       .Case("aes", true)
       .Case("pclmul", true)
       .Case("avx512vl", true)
       .Case("avx512bw", true)
       .Case("avx512dq", true)
       .Case("avx512cd", true)
+      .Case("avx512_vpopcntdq", true)
       .Case("avx512er", true)
       .Case("avx512pf", true)
       .Case("avx512vbmi", true)
       .Case("avx512ifma", true)
       .Default(false);
 }
 
 bool

lib/CodeGen/CGBuiltin.cpp

enum X86Features {
       AVX512VL,
       AVX512BW,
       AVX512DQ,
       AVX512CD,
       AVX512ER,
       AVX512PF,
       AVX512VBMI,
       AVX512IFMA,
+      AVX512_VPOPCNTDQ,
       MAX
     };
 
-    X86Features Feature = StringSwitch<X86Features>(FeatureStr)
-                              .Case("cmov", X86Features::CMOV)
-                              .Case("mmx", X86Features::MMX)
-                              .Case("popcnt", X86Features::POPCNT)
-                              .Case("sse", X86Features::SSE)
-                              .Case("sse2", X86Features::SSE2)
-                              .Case("sse3", X86Features::SSE3)
-                              .Case("ssse3", X86Features::SSSE3)
-                              .Case("sse4.1", X86Features::SSE4_1)
-                              .Case("sse4.2", X86Features::SSE4_2)
-                              .Case("avx", X86Features::AVX)
-                              .Case("avx2", X86Features::AVX2)
-                              .Case("sse4a", X86Features::SSE4_A)
-                              .Case("fma4", X86Features::FMA4)
-                              .Case("xop", X86Features::XOP)
-                              .Case("fma", X86Features::FMA)
-                              .Case("avx512f", X86Features::AVX512F)
-                              .Case("bmi", X86Features::BMI)
-                              .Case("bmi2", X86Features::BMI2)
-                              .Case("aes", X86Features::AES)
-                              .Case("pclmul", X86Features::PCLMUL)
-                              .Case("avx512vl", X86Features::AVX512VL)
-                              .Case("avx512bw", X86Features::AVX512BW)
-                              .Case("avx512dq", X86Features::AVX512DQ)
-                              .Case("avx512cd", X86Features::AVX512CD)
-                              .Case("avx512er", X86Features::AVX512ER)
-                              .Case("avx512pf", X86Features::AVX512PF)
-                              .Case("avx512vbmi", X86Features::AVX512VBMI)
-                              .Case("avx512ifma", X86Features::AVX512IFMA)
-                              .Default(X86Features::MAX);
+    X86Features Feature =
+        StringSwitch<X86Features>(FeatureStr)
+            .Case("avx512_vpopcntdq", X86Features::AVX512_VPOPCNTDQ)
     assert(Feature != X86Features::MAX && "Invalid feature!");
 
     // Matching the struct layout from the compiler-rt/libgcc structure that is
     // filled in:
     // unsigned int __cpu_vendor;
     // unsigned int __cpu_type;
     // unsigned int __cpu_subtype;
     // unsigned int __cpu_features[1];
#undef INTRINSIC_X86_XSAVE_ID
 
   case X86::BI__builtin_ia32_storess128_mask:
   case X86::BI__builtin_ia32_storesd128_mask: {
     return EmitX86MaskedStore(*this, Ops, 16);
   }
 
   case X86::BI__builtin_ia32_cvtmask2b128:
   case X86::BI__builtin_ia32_cvtmask2b256:
   case X86::BI__builtin_ia32_cvtmask2b512:
   case X86::BI__builtin_ia32_cvtmask2w128:
   case X86::BI__builtin_ia32_cvtmask2w256:
   case X86::BI__builtin_ia32_cvtmask2w512:
   case X86::BI__builtin_ia32_cvtmask2d128:
   case X86::BI__builtin_ia32_cvtmask2d256:
   case X86::BI__builtin_ia32_cvtmask2d512:
   case X86::BI__builtin_ia32_cvtmask2q128:
   case X86::BI__builtin_ia32_cvtmask2q256:
   case X86::BI__builtin_ia32_cvtmask2q512:

lib/Headers/CMakeLists.txt

 set(files
   adxintrin.h
   altivec.h
   ammintrin.h
   arm_acle.h
   armintr.h
   avx2intrin.h
   avx512bwintrin.h
   avx512cdintrin.h
+  avx512_vpopcntdqintrin.h
   avx512dqintrin.h
   avx512erintrin.h
   avx512fintrin.h
   avx512ifmaintrin.h
   avx512ifmavlintrin.h
   avx512pfintrin.h
   avx512vbmiintrin.h
   avx512vbmivlintrin.h

lib/Headers/avx512_vpopcntdqintrin.h

+/*===------------- avx512_vpopcntdqintrin.h - AVX512_VPOPCNTDQ intrinsics
+ *------------------===
+ *
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef __IMMINTRIN_H
+#error                                                                         \
+    "Never use <avx512_vpopcntdqintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX512_VPOPCNTDQINTRIN_H
+#define __AVX512_VPOPCNTDQINTRIN_H
+
+/* Define the default attributes for the functions in this file. */
+#define __DEFAULT_FN_ATTRS                                                     \
+  __attribute__(                                                               \
+      (__always_inline__, __nodebug__, __target__("avx512_vpopcntdq")))
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_popcnt_epi64(__m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntq_512_mask(
+      (__v8di)__A, (__v8di)_mm512_setzero_si512(), (__mmask8)-1);
+}
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS
+_mm512_mask_popcnt_epi64(__m512i __W, __mmask8 __U, __m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntq_512_mask((__v8di)__A, (__v8di)__W,
+                                                   (__mmask8)__U);
+}
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS
+_mm512_maskz_popcnt_epi64(__mmask8 __U, __m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntq_512_mask(
+      (__v8di)__A, (__v8di)_mm512_setzero_si512(), (__mmask8)__U);
+}
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_popcnt_epi32(__m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntd_512_mask(
+      (__v16si)__A, (__v16si)_mm512_setzero_si512(), (__mmask16)-1);
+}
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS
+_mm512_mask_popcnt_epi32(__m512i __W, __mmask16 __U, __m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntd_512_mask((__v16si)__A, (__v16si)__W,
+                                                   (__mmask16)__U);
+}
+
+static __inline__ __m512i __DEFAULT_FN_ATTRS
+_mm512_maskz_popcnt_epi32(__mmask16 __U, __m512i __A) {
+  return (__m512i)__builtin_ia32_vpopcntd_512_mask(
+      (__v16si)__A, (__v16si)_mm512_setzero_si512(), (__mmask16)__U);
+}
+
+#undef __DEFAULT_FN_ATTRS
+
+#endif

lib/Headers/immintrin.h

 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__AVX512BW__)
 #include <avx512bwintrin.h>
 #endif
 
 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__AVX512CD__)
 #include <avx512cdintrin.h>
 #endif
 
+#if !defined(_MSC_VER) || __has_feature(modules) ||                            \
+    defined(__AVX512_VPOPCNTDQ__)
+#include <avx512_vpopcntdqintrin.h>
+#endif
+
 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__AVX512DQ__)
 #include <avx512dqintrin.h>
 #endif
 
 #if !defined(_MSC_VER) || __has_feature(modules) || \
     (defined(__AVX512VL__) && defined(__AVX512BW__))
 #include <avx512vlbwintrin.h>
 #endif

test/CodeGen/attr-target-x86.c

 // CHECK: panda{{.*}} #3
 // CHECK: bar{{.*}} #0
 // CHECK: qux{{.*}} #1
 // CHECK: qax{{.*}} #4
 // CHECK: qq{{.*}} #5
 // CHECK: lake{{.*}} #6
 // CHECK: #0 = {{.*}}"target-cpu"="x86-64" "target-features"="+fxsr,+mmx,+sse,+sse2,+x87"
 // CHECK: #1 = {{.*}}"target-cpu"="ivybridge" "target-features"="+aes,+avx,+cx16,+f16c,+fsgsbase,+fxsr,+mmx,+pclmul,+popcnt,+rdrnd,+sse,+sse2,+sse3,+sse4.1,+sse4.2,+ssse3,+x87,+xsave,+xsaveopt"
-// CHECK: #2 = {{.*}}"target-cpu"="x86-64" "target-features"="+fxsr,+mmx,+sse,+x87,-aes,-avx,-avx2,-avx512bw,-avx512cd,-avx512dq,-avx512er,-avx512f,-avx512ifma,-avx512pf,-avx512vbmi,-avx512vl,-f16c,-fma,-fma4,-pclmul,-sha,-sse2,-sse3,-sse4.1,-sse4.2,-sse4a,-ssse3,-xop,-xsave,-xsaveopt"
+// CHECK: #2 = {{.*}}"target-cpu"="x86-64" "target-features"="+fxsr,+mmx,+sse,+x87,-aes,-avx,-avx2,-avx512_vpopcntdq,-avx512bw,-avx512cd,-avx512dq,-avx512er,-avx512f,-avx512ifma,-avx512pf,-avx512vbmi,-avx512vl,-f16c,-fma,-fma4,-pclmul,-sha,-sse2,-sse3,-sse4.1,-sse4.2,-sse4a,-ssse3,-xop,-xsave,-xsaveopt"
 // CHECK: #3 = {{.*}}"target-cpu"="x86-64" "target-features"="+fxsr,+mmx,+popcnt,+sse,+sse2,+sse3,+sse4.1,+sse4.2,+ssse3,+x87"
 // CHECK: #4 = {{.*}}"target-cpu"="ivybridge" "target-features"="+avx,+cx16,+f16c,+fsgsbase,+fxsr,+mmx,+pclmul,+popcnt,+rdrnd,+sse,+sse2,+sse3,+sse4.1,+sse4.2,+ssse3,+x87,+xsave,+xsaveopt,-aes"
 // CHECK: #5 = {{.*}}"target-cpu"="x86-64" "target-features"="+fxsr,+sse,+sse2,+x87,-3dnow,-3dnowa,-mmx"
 // CHECK: #6 = {{.*}}"target-cpu"="lakemont" "target-features"="+mmx,+sse,+sse2"

test/CodeGen/avx512_vpopcntdqintrin.c

+// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512_vpopcntdq -emit-llvm -o - -Wall -Werror | FileCheck %s
+
+#include <immintrin.h>
+
+__m512i test_mm512_popcnt_epi64(__m512i __A) {
+  // CHECK-LABEL: @test_mm512_popcnt_epi64
+  // CHECK: @llvm.x86.avx512.mask.popcnt.q.512
+  return _mm512_popcnt_epi64(__A);
+}
+__m512i test_mm512_mask_popcnt_epi64(__m512i __W, __mmask8 __U, __m512i __A) {
+  // CHECK-LABEL: @test_mm512_mask_popcnt_epi64
+  // CHECK: @llvm.x86.avx512.mask.popcnt.q.512
+  return _mm512_mask_popcnt_epi64(__W, __U, __A);
+}
+__m512i test_mm512_maskz_popcnt_epi64(__mmask8 __U, __m512i __A) {
+  // CHECK-LABEL: @test_mm512_maskz_popcnt_epi64
+  // CHECK: @llvm.x86.avx512.mask.popcnt.q.512
+  return _mm512_maskz_popcnt_epi64(__U, __A);
+}
+__m512i test_mm512_popcnt_epi32(__m512i __A) {
+  // CHECK-LABEL: @test_mm512_popcnt_epi32
+  // CHECK: @llvm.x86.avx512.mask.popcnt.d.512
+  return _mm512_popcnt_epi32(__A);
+}
+__m512i test_mm512_mask_popcnt_epi32(__m512i __W, __mmask16 __U, __m512i __A) {
+  // CHECK-LABEL: @test_mm512_mask_popcnt_epi32
+  // CHECK: @llvm.x86.avx512.mask.popcnt.d.512
+  return _mm512_mask_popcnt_epi32(__W, __U, __A);
+}
+__m512i test_mm512_maskz_popcnt_epi32(__mmask16 __U, __m512i __A) {
+  // CHECK-LABEL: @test_mm512_maskz_popcnt_epi32
+  // CHECK: @llvm.x86.avx512.mask.popcnt.d.512
+  return _mm512_maskz_popcnt_epi32(__U, __A);
+}