r268721 - [OPENMP 4.0] Codegen for 'declare simd' directive.
Alexey Bataev via cfe-commits
cfe-commits at lists.llvm.org
Wed May 11 20:03:23 PDT 2016
Yes, mangled names for vector versions are added as attributes to the
original function.
Best regards,
Alexey Bataev
=============
Software Engineer
Intel Compiler Team
12.05.2016 6:00, Hal Finkel пишет:
> Hi Alexey,
>
> As I recall, Xinmin's RFC discussed encoding the various possible manglings for vector functions in some attribute. Is that what this does? It is difficult to tell from the test case what's actually happing here.
>
> -Hal
>
> ----- Original Message -----
>> From: "Alexey Bataev" <a.bataev at hotmail.com>
>> To: "Hal Finkel" <hfinkel at anl.gov>
>> Cc: cfe-commits at lists.llvm.org
>> Sent: Wednesday, May 11, 2016 9:38:27 PM
>> Subject: Re: r268721 - [OPENMP 4.0] Codegen for 'declare simd' directive.
>>
>> Hal,
>>
>> According to Xinmin's RFC the codegen for 'declare simd' in clang
>> includes only generation of special (Vector ABI) mangled names for
>> future vectorized versions of elemental functions. Actual vector
>> versions must be generated by a backend. Xinmin and his team are
>> working
>> on this.
>>
>> The mangling part was accepted by David Majnemer (that's why I had to
>> implement it in clang), but RFC for a backend pass for generation of
>> vector functions is still under review, I think
>>
>> Best regards,
>> Alexey Bataev
>> =============
>> Software Engineer
>> Intel Compiler Team
>>
>> 11.05.2016 20:46, Hal Finkel пишет:
>>> Hi Alexey,
>>>
>>> I'm a bit confused regarding what this patch does, in part because
>>> the test case does not test any of the function argument or return
>>> types. Does it generate these "simd" functions taking scalar
>>> arguments? I'd expect that what this does is that it generates
>>> functions taking vector arguments and producing vector results. It
>>> would then extract the scalar values from those arguments and feed
>>> them into the scalar code (replicated N times), and then insert
>>> the results into the vector return value. The SLP vectorizer will
>>> then vectorize the function if profitable.
>>>
>>> Thanks again,
>>> Hal
>>>
>>> ----- Original Message -----
>>>> From: "Alexey Bataev via cfe-commits" <cfe-commits at lists.llvm.org>
>>>> To: cfe-commits at lists.llvm.org
>>>> Sent: Friday, May 6, 2016 4:40:08 AM
>>>> Subject: r268721 - [OPENMP 4.0] Codegen for 'declare simd'
>>>> directive.
>>>>
>>>> Author: abataev
>>>> Date: Fri May 6 04:40:08 2016
>>>> New Revision: 268721
>>>>
>>>> URL: http://llvm.org/viewvc/llvm-project?rev=268721&view=rev
>>>> Log:
>>>> [OPENMP 4.0] Codegen for 'declare simd' directive.
>>>>
>>>> OpenMP 4.0 adds support for elemental functions using declarative
>>>> directive '#pragma omp declare simd'. Patch adds mangling for simd
>>>> functions in accordance with
>>>> https://sourceware.org/glibc/wiki/libmvec?action=AttachFile&do=view&target=VectorABI.txt
>>>>
>>>> Added:
>>>> cfe/trunk/test/OpenMP/declare_simd_codegen.cpp
>>>> Modified:
>>>> cfe/trunk/lib/CodeGen/CGOpenMPRuntime.cpp
>>>> cfe/trunk/lib/CodeGen/CGOpenMPRuntime.h
>>>> cfe/trunk/lib/CodeGen/CodeGenFunction.cpp
>>>>
>>>> Modified: cfe/trunk/lib/CodeGen/CGOpenMPRuntime.cpp
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CGOpenMPRuntime.cpp?rev=268721&r1=268720&r2=268721&view=diff
>>>> ==============================================================================
>>>> --- cfe/trunk/lib/CodeGen/CGOpenMPRuntime.cpp (original)
>>>> +++ cfe/trunk/lib/CodeGen/CGOpenMPRuntime.cpp Fri May 6 04:40:08
>>>> 2016
>>>> @@ -6042,3 +6042,228 @@ void CGOpenMPRuntime::emitTargetEnterOrE
>>>> ThenGenRCG(CGF);
>>>> }
>>>> }
>>>> +
>>>> +namespace {
>>>> + /// Kind of parameter in a function with 'declare simd'
>>>> directive.
>>>> + enum ParamKindTy { LinearWithVarStride, Linear, Uniform, Vector
>>>> };
>>>> + /// Attribute set of the parameter.
>>>> + struct ParamAttrTy {
>>>> + ParamKindTy Kind = Vector;
>>>> + llvm::APSInt StrideOrArg;
>>>> + llvm::APSInt Alignment;
>>>> + };
>>>> +} // namespace
>>>> +
>>>> +static unsigned evaluateCDTSize(const FunctionDecl *FD,
>>>> + ArrayRef<ParamAttrTy> ParamAttrs)
>>>> {
>>>> + // Every vector variant of a SIMD-enabled function has a vector
>>>> length (VLEN).
>>>> + // If OpenMP clause "simdlen" is used, the VLEN is the value of
>>>> the argument
>>>> + // of that clause. The VLEN value must be power of 2.
>>>> + // In other case the notion of the function`s "characteristic
>>>> data
>>>> type" (CDT)
>>>> + // is used to compute the vector length.
>>>> + // CDT is defined in the following order:
>>>> + // a) For non-void function, the CDT is the return type.
>>>> + // b) If the function has any non-uniform, non-linear
>>>> parameters, then the
>>>> + // CDT is the type of the first such parameter.
>>>> + // c) If the CDT determined by a) or b) above is struct,
>>>> union,
>>>> or class
>>>> + // type which is pass-by-value (except for the type that maps
>>>> to
>>>> the
>>>> + // built-in complex data type), the characteristic data type
>>>> is
>>>> int.
>>>> + // d) If none of the above three cases is applicable, the CDT
>>>> is
>>>> int.
>>>> + // The VLEN is then determined based on the CDT and the size of
>>>> vector
>>>> + // register of that ISA for which current vector version is
>>>> generated. The
>>>> + // VLEN is computed using the formula below:
>>>> + // VLEN = sizeof(vector_register) / sizeof(CDT),
>>>> + // where vector register size specified in section 3.2.1
>>>> Registers
>>>> and the
>>>> + // Stack Frame of original AMD64 ABI document.
>>>> + QualType RetType = FD->getReturnType();
>>>> + if (RetType.isNull())
>>>> + return 0;
>>>> + ASTContext &C = FD->getASTContext();
>>>> + QualType CDT;
>>>> + if (!RetType.isNull() && !RetType->isVoidType())
>>>> + CDT = RetType;
>>>> + else {
>>>> + unsigned Offset = 0;
>>>> + if (auto *MD = dyn_cast<CXXMethodDecl>(FD)) {
>>>> + if (ParamAttrs[Offset].Kind == Vector)
>>>> + CDT = C.getPointerType(C.getRecordType(MD->getParent()));
>>>> + ++Offset;
>>>> + }
>>>> + if (CDT.isNull()) {
>>>> + for (unsigned I = 0, E = FD->getNumParams(); I < E; ++I) {
>>>> + if (ParamAttrs[I + Offset].Kind == Vector) {
>>>> + CDT = FD->getParamDecl(I)->getType();
>>>> + break;
>>>> + }
>>>> + }
>>>> + }
>>>> + }
>>>> + if (CDT.isNull())
>>>> + CDT = C.IntTy;
>>>> + CDT = CDT->getCanonicalTypeUnqualified();
>>>> + if (CDT->isRecordType() || CDT->isUnionType())
>>>> + CDT = C.IntTy;
>>>> + return C.getTypeSize(CDT);
>>>> +}
>>>> +
>>>> +static void
>>>> +emitX86DeclareSimdFunction(const FunctionDecl *FD, llvm::Function
>>>> *Fn,
>>>> + llvm::APSInt VLENVal,
>>>> + ArrayRef<ParamAttrTy> ParamAttrs,
>>>> + OMPDeclareSimdDeclAttr::BranchStateTy
>>>> State) {
>>>> + struct ISADataTy {
>>>> + char ISA;
>>>> + unsigned VecRegSize;
>>>> + };
>>>> + ISADataTy ISAData[] = {
>>>> + {
>>>> + 'b', 128
>>>> + }, // SSE
>>>> + {
>>>> + 'c', 256
>>>> + }, // AVX
>>>> + {
>>>> + 'd', 256
>>>> + }, // AVX2
>>>> + {
>>>> + 'e', 512
>>>> + }, // AVX512
>>>> + };
>>>> + llvm::SmallVector<char, 2> Masked;
>>>> + switch (State) {
>>>> + case OMPDeclareSimdDeclAttr::BS_Undefined:
>>>> + Masked.push_back('N');
>>>> + Masked.push_back('M');
>>>> + break;
>>>> + case OMPDeclareSimdDeclAttr::BS_Notinbranch:
>>>> + Masked.push_back('N');
>>>> + break;
>>>> + case OMPDeclareSimdDeclAttr::BS_Inbranch:
>>>> + Masked.push_back('M');
>>>> + break;
>>>> + }
>>>> + for (auto Mask : Masked) {
>>>> + for (auto &Data : ISAData) {
>>>> + SmallString<256> Buffer;
>>>> + llvm::raw_svector_ostream Out(Buffer);
>>>> + Out << "_ZGV" << Data.ISA << Mask;
>>>> + if (!VLENVal) {
>>>> + Out << llvm::APSInt::getUnsigned(Data.VecRegSize /
>>>> + evaluateCDTSize(FD,
>>>> ParamAttrs));
>>>> + } else
>>>> + Out << VLENVal;
>>>> + for (auto &ParamAttr : ParamAttrs) {
>>>> + switch (ParamAttr.Kind){
>>>> + case LinearWithVarStride:
>>>> + Out << 's' << ParamAttr.StrideOrArg;
>>>> + break;
>>>> + case Linear:
>>>> + Out << 'l';
>>>> + if (!!ParamAttr.StrideOrArg)
>>>> + Out << ParamAttr.StrideOrArg;
>>>> + break;
>>>> + case Uniform:
>>>> + Out << 'u';
>>>> + break;
>>>> + case Vector:
>>>> + Out << 'v';
>>>> + break;
>>>> + }
>>>> + if (!!ParamAttr.Alignment)
>>>> + Out << 'a' << ParamAttr.Alignment;
>>>> + }
>>>> + Out << '_' << Fn->getName();
>>>> + Fn->addFnAttr(Out.str());
>>>> + }
>>>> + }
>>>> +}
>>>> +
>>>> +void CGOpenMPRuntime::emitDeclareSimdFunction(const FunctionDecl
>>>> *FD,
>>>> + llvm::Function *Fn)
>>>> {
>>>> + ASTContext &C = CGM.getContext();
>>>> + FD = FD->getCanonicalDecl();
>>>> + // Map params to their positions in function decl.
>>>> + llvm::DenseMap<const Decl *, unsigned> ParamPositions;
>>>> + if (isa<CXXMethodDecl>(FD))
>>>> + ParamPositions.insert({FD, 0});
>>>> + unsigned ParamPos = ParamPositions.size();
>>>> + for (auto *P : FD->params()) {
>>>> + ParamPositions.insert({P->getCanonicalDecl(), ParamPos});
>>>> + ++ParamPos;
>>>> + }
>>>> + for (auto *Attr : FD->specific_attrs<OMPDeclareSimdDeclAttr>())
>>>> {
>>>> + llvm::SmallVector<ParamAttrTy, 8>
>>>> ParamAttrs(ParamPositions.size());
>>>> + // Mark uniform parameters.
>>>> + for (auto *E : Attr->uniforms()) {
>>>> + E = E->IgnoreParenImpCasts();
>>>> + unsigned Pos;
>>>> + if (isa<CXXThisExpr>(E))
>>>> + Pos = ParamPositions[FD];
>>>> + else {
>>>> + auto *PVD =
>>>> cast<ParmVarDecl>(cast<DeclRefExpr>(E)->getDecl())
>>>> + ->getCanonicalDecl();
>>>> + Pos = ParamPositions[PVD];
>>>> + }
>>>> + ParamAttrs[Pos].Kind = Uniform;
>>>> + }
>>>> + // Get alignment info.
>>>> + auto NI = Attr->alignments_begin();
>>>> + for (auto *E : Attr->aligneds()) {
>>>> + E = E->IgnoreParenImpCasts();
>>>> + unsigned Pos;
>>>> + QualType ParmTy;
>>>> + if (isa<CXXThisExpr>(E)) {
>>>> + Pos = ParamPositions[FD];
>>>> + ParmTy = E->getType();
>>>> + } else {
>>>> + auto *PVD =
>>>> cast<ParmVarDecl>(cast<DeclRefExpr>(E)->getDecl())
>>>> + ->getCanonicalDecl();
>>>> + Pos = ParamPositions[PVD];
>>>> + ParmTy = PVD->getType();
>>>> + }
>>>> + ParamAttrs[Pos].Alignment =
>>>> + (*NI) ? (*NI)->EvaluateKnownConstInt(C)
>>>> + : llvm::APSInt::getUnsigned(
>>>> +
>>>> C.toCharUnitsFromBits(C.getOpenMPDefaultSimdAlign(ParmTy))
>>>> + .getQuantity());
>>>> + ++NI;
>>>> + }
>>>> + // Mark linear parameters.
>>>> + auto SI = Attr->steps_begin();
>>>> + auto MI = Attr->modifiers_begin();
>>>> + for (auto *E : Attr->linears()) {
>>>> + E = E->IgnoreParenImpCasts();
>>>> + unsigned Pos;
>>>> + if (isa<CXXThisExpr>(E))
>>>> + Pos = ParamPositions[FD];
>>>> + else {
>>>> + auto *PVD =
>>>> cast<ParmVarDecl>(cast<DeclRefExpr>(E)->getDecl())
>>>> + ->getCanonicalDecl();
>>>> + Pos = ParamPositions[PVD];
>>>> + }
>>>> + auto &ParamAttr = ParamAttrs[Pos];
>>>> + ParamAttr.Kind = Linear;
>>>> + if (*SI) {
>>>> + if (!(*SI)->EvaluateAsInt(ParamAttr.StrideOrArg, C,
>>>> + Expr::SE_AllowSideEffects)) {
>>>> + if (auto *DRE =
>>>> cast<DeclRefExpr>((*SI)->IgnoreParenImpCasts())) {
>>>> + if (auto *StridePVD =
>>>> cast<ParmVarDecl>(DRE->getDecl()))
>>>> {
>>>> + ParamAttr.Kind = LinearWithVarStride;
>>>> + ParamAttr.StrideOrArg = llvm::APSInt::getUnsigned(
>>>> + ParamPositions[StridePVD->getCanonicalDecl()]);
>>>> + }
>>>> + }
>>>> + }
>>>> + }
>>>> + ++SI;
>>>> + ++MI;
>>>> + }
>>>> + llvm::APSInt VLENVal;
>>>> + if (const Expr *VLEN = Attr->getSimdlen())
>>>> + VLENVal = VLEN->EvaluateKnownConstInt(C);
>>>> + OMPDeclareSimdDeclAttr::BranchStateTy State =
>>>> Attr->getBranchState();
>>>> + if (CGM.getTriple().getArch() == llvm::Triple::x86 ||
>>>> + CGM.getTriple().getArch() == llvm::Triple::x86_64)
>>>> + emitX86DeclareSimdFunction(FD, Fn, VLENVal, ParamAttrs,
>>>> State);
>>>> + }
>>>> +}
>>>>
>>>> Modified: cfe/trunk/lib/CodeGen/CGOpenMPRuntime.h
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CGOpenMPRuntime.h?rev=268721&r1=268720&r2=268721&view=diff
>>>> ==============================================================================
>>>> --- cfe/trunk/lib/CodeGen/CGOpenMPRuntime.h (original)
>>>> +++ cfe/trunk/lib/CodeGen/CGOpenMPRuntime.h Fri May 6 04:40:08
>>>> 2016
>>>> @@ -1013,6 +1013,13 @@ public:
>>>> const
>>>> OMPExecutableDirective
>>>> &D,
>>>> const Expr *IfCond,
>>>> const Expr
>>>> *Device);
>>>> +
>>>> + /// Marks function \a Fn with properly mangled versions of
>>>> vector
>>>> functions.
>>>> + /// \param FD Function marked as 'declare simd'.
>>>> + /// \param Fn LLVM function that must be marked with 'declare
>>>> simd'
>>>> + /// attributes.
>>>> + virtual void emitDeclareSimdFunction(const FunctionDecl *FD,
>>>> + llvm::Function *Fn);
>>>> };
>>>>
>>>> } // namespace CodeGen
>>>>
>>>> Modified: cfe/trunk/lib/CodeGen/CodeGenFunction.cpp
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CodeGenFunction.cpp?rev=268721&r1=268720&r2=268721&view=diff
>>>> ==============================================================================
>>>> --- cfe/trunk/lib/CodeGen/CodeGenFunction.cpp (original)
>>>> +++ cfe/trunk/lib/CodeGen/CodeGenFunction.cpp Fri May 6 04:40:08
>>>> 2016
>>>> @@ -710,6 +710,8 @@ void CodeGenFunction::StartFunction(Glob
>>>> }
>>>> } else if (!FD->hasAttr<AlwaysInlineAttr>())
>>>> Fn->addFnAttr(llvm::Attribute::NoInline);
>>>> + if (CGM.getLangOpts().OpenMP &&
>>>> FD->hasAttr<OMPDeclareSimdDeclAttr>())
>>>> + CGM.getOpenMPRuntime().emitDeclareSimdFunction(FD, Fn);
>>>> }
>>>>
>>>> // Add no-jump-tables value.
>>>>
>>>> Added: cfe/trunk/test/OpenMP/declare_simd_codegen.cpp
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/cfe/trunk/test/OpenMP/declare_simd_codegen.cpp?rev=268721&view=auto
>>>> ==============================================================================
>>>> --- cfe/trunk/test/OpenMP/declare_simd_codegen.cpp (added)
>>>> +++ cfe/trunk/test/OpenMP/declare_simd_codegen.cpp Fri May 6
>>>> 04:40:08 2016
>>>> @@ -0,0 +1,288 @@
>>>> +// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp
>>>> -x
>>>> c++ -emit-llvm %s -o - -femit-all-decls | FileCheck %s
>>>> +// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10
>>>> -emit-pch -o %t %s
>>>> +// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10
>>>> -include-pch %t -verify %s -emit-llvm -o - -femit-all-decls |
>>>> FileCheck %s
>>>> +// expected-no-diagnostics
>>>> +// REQUIRES: x86-registered-target
>>>> +#ifndef HEADER
>>>> +#define HEADER
>>>> +
>>>> +#pragma omp declare simd linear(d : 8)
>>>> +#pragma omp declare simd inbranch simdlen(32)
>>>> +#pragma omp declare simd notinbranch
>>>> +void add_1(float *d) {}
>>>> +
>>>> +#pragma omp declare simd aligned(hp, hp2)
>>>> +template <class C>
>>>> +void h(C *hp, C *hp2, C *hq, C *lin) {
>>>> +}
>>>> +
>>>> +// Explicit specialization with <C=int>.
>>>> +// Pragmas need to be same, otherwise standard says that's
>>>> undefined
>>>> behavior.
>>>> +#pragma omp declare simd aligned(hp, hp2)
>>>> +template <>
>>>> +void h(int *hp, int *hp2, int *hq, int *lin) {
>>>> + // Implicit specialization with <C=float>.
>>>> + // This is special case where the directive is stored by Sema
>>>> and
>>>> is
>>>> + // generated together with the (pending) function instatiation.
>>>> + h((float *)hp, (float *)hp2, (float *)hq, (float *)lin);
>>>> +}
>>>> +
>>>> +class VV {
>>>> +public:
>>>> +#pragma omp declare simd uniform(this, a) linear(val(b) : a)
>>>> + int add(int a, int b) __attribute__((cold)) { return a + b; }
>>>> +
>>>> +#pragma omp declare simd aligned(b : 4) aligned(a) linear(ref(b)
>>>> :
>>>> 4) linear(this, a)
>>>> + float taddpf(float *a, float *&b) { return *a + *b; }
>>>> +
>>>> +#pragma omp declare simd linear(uval(c) : 8)
>>>> +#pragma omp declare simd aligned(b : 8)
>>>> + int tadd(int (&b)[], int &c) { return x[b[0]] + b[0]; }
>>>> +
>>>> +private:
>>>> + int x[10];
>>>> +} vv;
>>>> +
>>>> +template <int X, typename T>
>>>> +class TVV {
>>>> +public:
>>>> +#pragma omp declare simd simdlen(X)
>>>> + int tadd(int a, int b) { return a + b; }
>>>> +
>>>> +#pragma omp declare simd aligned(a : X * 2) aligned(b)
>>>> linear(ref(b)
>>>> : X)
>>>> + float taddpf(float *a, T *&b) { return *a + *b; }
>>>> +
>>>> +#pragma omp declare simd
>>>> +#pragma omp declare simd uniform(this, b)
>>>> + int tadd(int b) { return x[b] + b; }
>>>> +
>>>> +private:
>>>> + int x[X];
>>>> +};
>>>> +
>>>> +#pragma omp declare simd simdlen(N) aligned(b : N * 2)
>>>> linear(uval(c) : N)
>>>> +template <int N>
>>>> +void foo(int (&b)[N], float *&c) {}
>>>> +
>>>> +TVV<16, float> t16;
>>>> +
>>>> +void f(int (&g)[]) {
>>>> + float a = 1.0f, b = 2.0f;
>>>> + float *p = &b;
>>>> + float r = t16.taddpf(&a, p);
>>>> + int res = t16.tadd(b);
>>>> + int c[64];
>>>> + vv.add(res, res);
>>>> + vv.taddpf(p, p);
>>>> + vv.tadd(g, res);
>>>> + foo(c, p);
>>>> +}
>>>> +
>>>> +#pragma omp declare simd
>>>> +#pragma omp declare simd notinbranch aligned(a : 32)
>>>> +int bar(VV v, float *a) { return 0; }
>>>> +#pragma omp declare simd
>>>> +#pragma omp declare simd notinbranch aligned(a)
>>>> +float baz(VV v, int a[]) { return 0; }
>>>> +#pragma omp declare simd
>>>> +#pragma omp declare simd notinbranch aligned(a)
>>>> +double bay(VV v, double *&a) { return 0; }
>>>> +#pragma omp declare simd
>>>> +#pragma omp declare simd inbranch linear(a : b) uniform(v, b)
>>>> +void bax(VV v, double *a, int b) {}
>>>> +#pragma omp declare simd uniform(q) aligned(q : 16) linear(k : 1)
>>>> +float foo(float *q, float x, int k) { return 0; }
>>>> +#pragma omp declare simd notinbranch
>>>> +double foo(double x) { return 0; }
>>>> +
>>>> +// CHECK-DAG: define {{.+}}@_Z5add_1Pf(
>>>> +// CHECK-DAG: define {{.+}}@_Z1hIiEvPT_S1_S1_S1_(
>>>> +// CHECK-DAG: define {{.+}}@_Z1hIfEvPT_S1_S1_S1_(
>>>> +// CHECK-DAG: define {{.+}}@_ZN2VV3addEii(
>>>> +// CHECK-DAG: define {{.+}}@_ZN2VV6taddpfEPfRS0_(
>>>> +// CHECK-DAG: define {{.+}}@_ZN2VV4taddERA_iRi(
>>>> +// CHECK-DAG: define {{.+}}@_Z1fRA_i(
>>>> +// CHECK-DAG: define {{.+}}@_ZN3TVVILi16EfE6taddpfEPfRS1_(
>>>> +// CHECK-DAG: define {{.+}}@_ZN3TVVILi16EfE4taddEi(
>>>> +// CHECK-DAG: define {{.+}}@_Z3fooILi64EEvRAT__iRPf(
>>>> +// CHECK-DAG: define {{.+}}@_Z3bar2VVPf(
>>>> +// CHECK-DAG: define {{.+}}@_Z3baz2VVPi(
>>>> +// CHECK-DAG: define {{.+}}@_Z3bay2VVRPd(
>>>> +// CHECK-DAG: define {{.+}}@_Z3bax2VVPdi(
>>>> +// CHECK-DAG: define {{.+}}@_Z3fooPffi(
>>>> +// CHECK-DAG: define {{.+}}@_Z3food(
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVbN4l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVcM8l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVcN8l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVdM8l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVdN8l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVeM16l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVeN16l8__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVbM32v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVcM32v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVdM32v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVeM32v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVbN2v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVcN4v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVdN4v__Z5add_1Pf"
>>>> +// CHECK-DAG: "_ZGVeN8v__Z5add_1Pf"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM2va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVbN2va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVcM4va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVcN4va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVdM4va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVdN4va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVeM8va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVeN8va16va16vv__Z1hIiEvPT_S1_S1_S1_"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM2va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVbN2va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVcM4va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVcN4va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVdM4va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVdN4va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVeM8va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +// CHECK-DAG: "_ZGVeN8va16va16vv__Z1hIfEvPT_S1_S1_S1_"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVbN4uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVcM8uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVcN8uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVdM8uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVdN8uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVeM16uus1__ZN2VV3addEii"
>>>> +// CHECK-DAG: "_ZGVeN16uus1__ZN2VV3addEii"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVbN4lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVcM8lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVcN8lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVdM8lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVdN8lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVeM16lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +// CHECK-DAG: "_ZGVeN16lla16l4a4__ZN2VV6taddpfEPfRS0_"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVbN4vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVcM8vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVcN8vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVdM8vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVdN8vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVeM16vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVeN16vvl8__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVbM4vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVbN4vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVcM8vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVcN8vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVdM8vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVdN8vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVeM16vva8v__ZN2VV4taddERA_iRi"
>>>> +// CHECK-DAG: "_ZGVeN16vva8v__ZN2VV4taddERA_iRi"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVbN4vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVcM8vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVcN8vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVdM8vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVdN8vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVeM16vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +// CHECK-DAG: "_ZGVeN16vva32l16a16__ZN3TVVILi16EfE6taddpfEPfRS1_"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVbN4uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVcM8uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVcN8uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVdM8uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVdN8uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVeM16uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVeN16uu__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVbM4vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVbN4vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVcM8vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVcN8vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVdM8vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVdN8vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVeM16vv__ZN3TVVILi16EfE4taddEi"
>>>> +// CHECK-DAG: "_ZGVeN16vv__ZN3TVVILi16EfE4taddEi"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVbN64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVcM64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVcN64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVdM64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVdN64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVeM64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +// CHECK-DAG: "_ZGVeN64va128l64__Z3fooILi64EEvRAT__iRPf"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVbN4vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVcM8vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVcN8vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVdM8vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVdN8vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVeM16vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVeN16vv__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVbN4vva32__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVcN8vva32__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVdN8vva32__Z3bar2VVPf"
>>>> +// CHECK-DAG: "_ZGVeN16vva32__Z3bar2VVPf"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVbN4vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVcM8vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVcN8vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVdM8vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVdN8vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVeM16vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVeN16vv__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVbN4vva16__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVcN8vva16__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVdN8vva16__Z3baz2VVPi"
>>>> +// CHECK-DAG: "_ZGVeN16vva16__Z3baz2VVPi"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM2vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVbN2vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVcM4vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVcN4vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVdM4vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVdN4vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVeM8vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVeN8vv__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVbN2vva16__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVcN4vva16__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVdN4vva16__Z3bay2VVRPd"
>>>> +// CHECK-DAG: "_ZGVeN8vva16__Z3bay2VVRPd"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4us2u__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVcM8us2u__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVdM8us2u__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVeM16us2u__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVbM4vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVbN4vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVcM8vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVcN8vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVdM8vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVdN8vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVeM16vvv__Z3bax2VVPdi"
>>>> +// CHECK-DAG: "_ZGVeN16vvv__Z3bax2VVPdi"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbM4ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVbN4ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVcM8ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVcN8ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVdM8ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVdN8ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVeM16ua16vl1__Z3fooPffi"
>>>> +// CHECK-DAG: "_ZGVeN16ua16vl1__Z3fooPffi"
>>>> +
>>>> +// CHECK-DAG: "_ZGVbN2v__Z3food"
>>>> +// CHECK-DAG: "_ZGVcN4v__Z3food"
>>>> +// CHECK-DAG: "_ZGVdN4v__Z3food"
>>>> +// CHECK-DAG: "_ZGVeN8v__Z3food"
>>>> +
>>>> +// CHECK-NOT: "_ZGV{{.+}}__Z1fRA_i
>>>> +
>>>> +#endif
>>>>
>>>>
>>>> _______________________________________________
>>>> cfe-commits mailing list
>>>> cfe-commits at lists.llvm.org
>>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
>>>>
>>
More information about the cfe-commits
mailing list