https://github.com/kimsh02 updated
https://github.com/llvm/llvm-project/pull/158703
>From 1b81bf79bc7e6a2c1df846b27e4a95cf1cb97090 Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Mon, 15 Sep 2025 10:58:34 -0700
Subject: [PATCH 01/14] [Clang] VectorExprEvaluator::VisitCallExpr /
InterpretBuiltin - add AVX512 VPTERNLOGD/VPTERNLOGQ intrinsics to be used in
constexpr
---
clang/include/clang/Basic/BuiltinsX86.td | 16 +-
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 64 ++++-
clang/lib/AST/ExprConstant.cpp | 91 +++++++
clang/test/CodeGen/X86/avx512f-builtins.c | 138 ++++++++++
clang/test/CodeGen/X86/avx512vl-builtins.c | 277 +++++++++++++++++++++
5 files changed, 580 insertions(+), 6 deletions(-)
diff --git a/clang/include/clang/Basic/BuiltinsX86.td
b/clang/include/clang/Basic/BuiltinsX86.td
index 044c755d4d7cf..b411e583b1a2c 100644
--- a/clang/include/clang/Basic/BuiltinsX86.td
+++ b/clang/include/clang/Basic/BuiltinsX86.td
@@ -2402,28 +2402,36 @@ let Features = "avx512f", Attributes = [NoThrow, Const,
RequiredVectorWidth<512>
def psraq512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long
int>, _Vector<2, long long int>)">;
def psrld512 : X86Builtin<"_Vector<16, int>(_Vector<16, int>, _Vector<4,
int>)">;
def psrlq512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long
int>, _Vector<2, long long int>)">;
+}
+
+let Features = "avx512f",
+ Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in {
def pternlogd512_mask : X86Builtin<"_Vector<16, int>(_Vector<16, int>,
_Vector<16, int>, _Vector<16, int>, _Constant int, unsigned short)">;
def pternlogd512_maskz : X86Builtin<"_Vector<16, int>(_Vector<16, int>,
_Vector<16, int>, _Vector<16, int>, _Constant int, unsigned short)">;
def pternlogq512_mask : X86Builtin<"_Vector<8, long long int>(_Vector<8,
long long int>, _Vector<8, long long int>, _Vector<8, long long int>, _Constant
int, unsigned char)">;
def pternlogq512_maskz : X86Builtin<"_Vector<8, long long int>(_Vector<8,
long long int>, _Vector<8, long long int>, _Vector<8, long long int>, _Constant
int, unsigned char)">;
}
-let Features = "avx512vl", Attributes = [NoThrow, Const,
RequiredVectorWidth<128>] in {
+let Features = "avx512vl",
+ Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
def pternlogd128_mask : X86Builtin<"_Vector<4, int>(_Vector<4, int>,
_Vector<4, int>, _Vector<4, int>, _Constant int, unsigned char)">;
def pternlogd128_maskz : X86Builtin<"_Vector<4, int>(_Vector<4, int>,
_Vector<4, int>, _Vector<4, int>, _Constant int, unsigned char)">;
}
-let Features = "avx512vl", Attributes = [NoThrow, Const,
RequiredVectorWidth<256>] in {
+let Features = "avx512vl",
+ Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in {
def pternlogd256_mask : X86Builtin<"_Vector<8, int>(_Vector<8, int>,
_Vector<8, int>, _Vector<8, int>, _Constant int, unsigned char)">;
def pternlogd256_maskz : X86Builtin<"_Vector<8, int>(_Vector<8, int>,
_Vector<8, int>, _Vector<8, int>, _Constant int, unsigned char)">;
}
-let Features = "avx512vl", Attributes = [NoThrow, Const,
RequiredVectorWidth<128>] in {
+let Features = "avx512vl",
+ Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
def pternlogq128_mask : X86Builtin<"_Vector<2, long long int>(_Vector<2,
long long int>, _Vector<2, long long int>, _Vector<2, long long int>, _Constant
int, unsigned char)">;
def pternlogq128_maskz : X86Builtin<"_Vector<2, long long int>(_Vector<2,
long long int>, _Vector<2, long long int>, _Vector<2, long long int>, _Constant
int, unsigned char)">;
}
-let Features = "avx512vl", Attributes = [NoThrow, Const,
RequiredVectorWidth<256>] in {
+let Features = "avx512vl",
+ Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in {
def pternlogq256_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4,
long long int>, _Vector<4, long long int>, _Vector<4, long long int>, _Constant
int, unsigned char)">;
def pternlogq256_maskz : X86Builtin<"_Vector<4, long long int>(_Vector<4,
long long int>, _Vector<4, long long int>, _Vector<4, long long int>, _Constant
int, unsigned char)">;
}
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 77729a5d67c87..00466aa7ae663 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2952,10 +2952,57 @@ static bool
interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC,
});
Dst.initializeAllElements();
-
return true;
}
+static bool interp__builtin_pternlog(InterpState &S, CodePtr OpPC,
+ const CallExpr *Call, bool MaskZ) {
+ assert(Call->getNumArgs() == 5);
+
+ const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
+ unsigned DstLen = VecT->getNumElements();
+ PrimType DstElemT = *S.getContext().classify(VecT->getElementType());
+
+ APSInt U = popToAPSInt(S.Stk, *S.getContext().classify(Call->getArg(4)));
+ APSInt Imm = popToAPSInt(S.Stk, *S.getContext().classify(Call->getArg(3)));
+ const Pointer &C = S.Stk.pop<Pointer>();
+ const Pointer &B = S.Stk.pop<Pointer>();
+ const Pointer &A = S.Stk.pop<Pointer>();
+
+ const Pointer &Dst = S.Stk.peek<Pointer>();
+
+ for (unsigned I = 0; I < DstLen; ++I) {
+ APSInt a, b, c;
+ INT_TYPE_SWITCH(DstElemT, {
+ a = A.elem<T>(I).toAPSInt();
+ b = B.elem<T>(I).toAPSInt();
+ c = C.elem<T>(I).toAPSInt();
+ });
+
+ unsigned BitWidth = a.getBitWidth();
+ APInt R(BitWidth, 0);
+ bool DstUnsigned = a.isUnsigned();
+
+ if (U[I]) {
+ for (unsigned Bit = 0; Bit < BitWidth; ++Bit) {
+ unsigned Idx = (a[Bit] << 2) | (b[Bit] << 1) | (c[Bit]);
+ R.setBitVal(Bit, Imm[Idx]);
+ }
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
+ Dst.elem<T>(I) = static_cast<T>(APSInt(R, DstUnsigned));
+ });
+ } else if (MaskZ) {
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
+ Dst.elem<T>(I) = static_cast<T>(APSInt(R, DstUnsigned));
+ });
+ } else {
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT,
+ { Dst.elem<T>(I) = static_cast<T>(a); });
+ }
+ }
+ Dst.initializeAllElements();
+}
+
bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
uint32_t BuiltinID) {
if (!S.getASTContext().BuiltinInfo.isConstantEvaluated(BuiltinID))
@@ -3606,7 +3653,20 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC,
const CallExpr *Call,
case X86::BI__builtin_ia32_selectpd_256:
case X86::BI__builtin_ia32_selectpd_512:
return interp__builtin_select(S, OpPC, Call);
-
+ case X86::BI__builtin_ia32_pternlogd128_mask:
+ case X86::BI__builtin_ia32_pternlogd256_mask:
+ case X86::BI__builtin_ia32_pternlogd512_mask:
+ case X86::BI__builtin_ia32_pternlogq128_mask:
+ case X86::BI__builtin_ia32_pternlogq256_mask:
+ case X86::BI__builtin_ia32_pternlogq512_mask:
+ return interp__builtin_pternlog(S, OpPC, Call, false);
+ case X86::BI__builtin_ia32_pternlogd128_maskz:
+ case X86::BI__builtin_ia32_pternlogd256_maskz:
+ case X86::BI__builtin_ia32_pternlogd512_maskz:
+ case X86::BI__builtin_ia32_pternlogq128_maskz:
+ case X86::BI__builtin_ia32_pternlogq256_maskz:
+ case X86::BI__builtin_ia32_pternlogq512_maskz:
+ return interp__builtin_pternlog(S, OpPC, Call, true);
case Builtin::BI__builtin_elementwise_fshl:
return interp__builtin_elementwise_triop(S, OpPC, Call,
llvm::APIntOps::fshl);
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index 3b9ca82910033..db725ba935b99 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -12025,6 +12025,97 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr
*E) {
return Success(APValue(ResultElements.data(), ResultElements.size()), E);
}
+ case X86::BI__builtin_ia32_pternlogd128_mask:
+ case X86::BI__builtin_ia32_pternlogd256_mask:
+ case X86::BI__builtin_ia32_pternlogd512_mask:
+ case X86::BI__builtin_ia32_pternlogq128_mask:
+ case X86::BI__builtin_ia32_pternlogq256_mask:
+ case X86::BI__builtin_ia32_pternlogq512_mask: {
+ APValue AValue, BValue, CValue, ImmValue, UValue;
+ if (!EvaluateAsRValue(Info, E->getArg(0), AValue) ||
+ !EvaluateAsRValue(Info, E->getArg(1), BValue) ||
+ !EvaluateAsRValue(Info, E->getArg(2), CValue) ||
+ !EvaluateAsRValue(Info, E->getArg(3), ImmValue) ||
+ !EvaluateAsRValue(Info, E->getArg(4), UValue))
+ return false;
+
+ QualType DestEltTy = E->getType()->castAs<VectorType>()->getElementType();
+ bool DestUnsigned = DestEltTy->isUnsignedIntegerOrEnumerationType();
+ APInt Imm = ImmValue.getInt();
+ APInt U = UValue.getInt();
+ unsigned ResultLen = AValue.getVectorLength();
+ SmallVector<APValue, 16> ResultElements;
+ ResultElements.reserve(ResultLen);
+
+ for (unsigned EltNum = 0; EltNum < ResultLen; ++EltNum) {
+ APInt ALane = AValue.getVectorElt(EltNum).getInt();
+ APInt BLane = BValue.getVectorElt(EltNum).getInt();
+ APInt CLane = CValue.getVectorElt(EltNum).getInt();
+
+ if (U[EltNum]) {
+ unsigned BitWidth = ALane.getBitWidth();
+ APInt ResLane(BitWidth, 0);
+
+ for (unsigned Bit = 0; Bit < BitWidth; ++Bit) {
+ unsigned ABit = ALane[Bit];
+ unsigned BBit = BLane[Bit];
+ unsigned CBit = CLane[Bit];
+
+ unsigned Idx = (ABit << 2) | (BBit << 1) | CBit;
+ ResLane.setBitVal(Bit, Imm[Idx]);
+ }
+ ResultElements.push_back(APValue(APSInt(ResLane, DestUnsigned)));
+ } else {
+ ResultElements.push_back(APValue(APSInt(ALane, DestUnsigned)));
+ }
+ }
+ return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+ }
+ case X86::BI__builtin_ia32_pternlogd128_maskz:
+ case X86::BI__builtin_ia32_pternlogd256_maskz:
+ case X86::BI__builtin_ia32_pternlogd512_maskz:
+ case X86::BI__builtin_ia32_pternlogq128_maskz:
+ case X86::BI__builtin_ia32_pternlogq256_maskz:
+ case X86::BI__builtin_ia32_pternlogq512_maskz: {
+ APValue AValue, BValue, CValue, ImmValue, UValue;
+ if (!EvaluateAsRValue(Info, E->getArg(0), AValue) ||
+ !EvaluateAsRValue(Info, E->getArg(1), BValue) ||
+ !EvaluateAsRValue(Info, E->getArg(2), CValue) ||
+ !EvaluateAsRValue(Info, E->getArg(3), ImmValue) ||
+ !EvaluateAsRValue(Info, E->getArg(4), UValue))
+ return false;
+
+ QualType DestEltTy = E->getType()->castAs<VectorType>()->getElementType();
+ bool DestUnsigned = DestEltTy->isUnsignedIntegerOrEnumerationType();
+ APInt Imm = ImmValue.getInt();
+ APInt U = UValue.getInt();
+ unsigned ResultLen = AValue.getVectorLength();
+ SmallVector<APValue, 16> ResultElements;
+ ResultElements.reserve(ResultLen);
+
+ for (unsigned EltNum = 0; EltNum < ResultLen; ++EltNum) {
+ APInt ALane = AValue.getVectorElt(EltNum).getInt();
+ APInt BLane = BValue.getVectorElt(EltNum).getInt();
+ APInt CLane = CValue.getVectorElt(EltNum).getInt();
+
+ unsigned BitWidth = ALane.getBitWidth();
+ APInt ResLane(BitWidth, 0);
+
+ if (U[EltNum]) {
+ for (unsigned Bit = 0; Bit < BitWidth; ++Bit) {
+ unsigned ABit = ALane[Bit];
+ unsigned BBit = BLane[Bit];
+ unsigned CBit = CLane[Bit];
+
+ unsigned Idx = (ABit << 2) | (BBit << 1) | CBit;
+ ResLane.setBitVal(Bit, Imm[Idx]);
+ }
+ }
+ ResultElements.push_back(APValue(APSInt(ResLane, DestUnsigned)));
+ }
+ return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+ }
+
case Builtin::BI__builtin_elementwise_clzg:
case Builtin::BI__builtin_elementwise_ctzg: {
APValue SourceLHS;
diff --git a/clang/test/CodeGen/X86/avx512f-builtins.c
b/clang/test/CodeGen/X86/avx512f-builtins.c
index e25eb9846431d..9fa450c68e503 100644
--- a/clang/test/CodeGen/X86/avx512f-builtins.c
+++ b/clang/test/CodeGen/X86/avx512f-builtins.c
@@ -6246,6 +6246,27 @@ __m512i test_mm512_ternarylogic_epi32(__m512i __A,
__m512i __B, __m512i __C) {
// CHECK: @llvm.x86.avx512.pternlog.d.512({{.*}}, i32 240)
return _mm512_ternarylogic_epi32(__A, __B, __C, _MM_TERNLOG_A);
}
+TEST_CONSTEXPR(match_v16si(
+ _mm512_ternarylogic_epi32(
+ ((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB,
0xC));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_ternarylogic_epi32(
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF,
0xF));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_ternarylogic_epi32(
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0));
__m512i test_mm512_mask_ternarylogic_epi32(__m512i __A, __mmask16 __U, __m512i
__B, __m512i __C) {
// CHECK-LABEL: test_mm512_mask_ternarylogic_epi32
@@ -6253,6 +6274,30 @@ __m512i test_mm512_mask_ternarylogic_epi32(__m512i __A,
__mmask16 __U, __m512i _
// CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}}
return _mm512_mask_ternarylogic_epi32(__A, __U, __B, __C, _MM_TERNLOG_B);
}
+TEST_CONSTEXPR(match_v16si(
+ _mm512_mask_ternarylogic_epi32(
+ ((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ (__mmask16)0x3333,
+ ((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0, 0xB, 0xC,
-0x1, 0x0));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_mask_ternarylogic_epi32(
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask16)0xCCCC,
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF,
0xF));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_mask_ternarylogic_epi32(
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask16)0x5555,
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0,
0x9));
__m512i test_mm512_maskz_ternarylogic_epi32(__mmask16 __U, __m512i __A,
__m512i __B, __m512i __C) {
// CHECK-LABEL: test_mm512_maskz_ternarylogic_epi32
@@ -6260,12 +6305,57 @@ __m512i test_mm512_maskz_ternarylogic_epi32(__mmask16
__U, __m512i __A, __m512i
// CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32>
zeroinitializer
return _mm512_maskz_ternarylogic_epi32(__U, __A, __B, __C, _MM_TERNLOG_C);
}
+TEST_CONSTEXPR(match_v16si(
+ _mm512_maskz_ternarylogic_epi32(
+ (__mmask16)0x3333,
+ ((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0,
0x0));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_maskz_ternarylogic_epi32(
+ (__mmask16)0xCCCC,
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF,
0xF));
+TEST_CONSTEXPR(match_v16si(
+ _mm512_maskz_ternarylogic_epi32(
+ (__mmask16)0x5555,
+ ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0));
__m512i test_mm512_ternarylogic_epi64(__m512i __A, __m512i __B, __m512i __C) {
// CHECK-LABEL: test_mm512_ternarylogic_epi64
// CHECK: @llvm.x86.avx512.pternlog.q.512({{.*}}, i32 192)
return _mm512_ternarylogic_epi64(__A, __B, __C, _MM_TERNLOG_A &
_MM_TERNLOG_B);
}
+TEST_CONSTEXPR(match_v8di(
+ _mm512_ternarylogic_epi64(
+ ((__m512i)((__v8di){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m512i)((__v8di){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v8di){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_ternarylogic_epi64(
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_ternarylogic_epi64(
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0));
__m512i test_mm512_mask_ternarylogic_epi64(__m512i __A, __mmask8 __U, __m512i
__B, __m512i __C) {
// CHECK-LABEL: test_mm512_mask_ternarylogic_epi64
@@ -6273,6 +6363,30 @@ __m512i test_mm512_mask_ternarylogic_epi64(__m512i __A,
__mmask8 __U, __m512i __
// CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}
return _mm512_mask_ternarylogic_epi64(__A, __U, __B, __C, _MM_TERNLOG_B |
_MM_TERNLOG_C);
}
+TEST_CONSTEXPR(match_v8di(
+ _mm512_mask_ternarylogic_epi64(
+ ((__m512i)((__v8di){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ (__mmask8)0x33,
+ ((__m512i)((__v8di){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v8di){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_mask_ternarylogic_epi64(
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0xCC,
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_mask_ternarylogic_epi64(
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0x55,
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9));
__m512i test_mm512_maskz_ternarylogic_epi64(__mmask8 __U, __m512i __A, __m512i
__B, __m512i __C) {
// CHECK-LABEL: test_mm512_maskz_ternarylogic_epi64
@@ -6280,6 +6394,30 @@ __m512i test_mm512_maskz_ternarylogic_epi64(__mmask8
__U, __m512i __A, __m512i _
// CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64>
zeroinitializer
return _mm512_maskz_ternarylogic_epi64(__U, __A, __B, __C, ~_MM_TERNLOG_A |
(_MM_TERNLOG_B ^ _MM_TERNLOG_C));
}
+TEST_CONSTEXPR(match_v8di(
+ _mm512_maskz_ternarylogic_epi64(
+ (__mmask8)0x33,
+ ((__m512i)((__v8di){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m512i)((__v8di){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m512i)((__v8di){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_maskz_ternarylogic_epi64(
+ (__mmask8)0xCC,
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8di(
+ _mm512_maskz_ternarylogic_epi64(
+ (__mmask8)0x55,
+ ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0));
__m512 test_mm512_shuffle_f32x4(__m512 __A, __m512 __B) {
// CHECK-LABEL: test_mm512_shuffle_f32x4
diff --git a/clang/test/CodeGen/X86/avx512vl-builtins.c
b/clang/test/CodeGen/X86/avx512vl-builtins.c
index 465c43dff8493..d29045af6e303 100644
--- a/clang/test/CodeGen/X86/avx512vl-builtins.c
+++ b/clang/test/CodeGen/X86/avx512vl-builtins.c
@@ -8328,6 +8328,27 @@ __m128i test_mm_ternarylogic_epi32(__m128i __A, __m128i
__B, __m128i __C) {
// CHECK: @llvm.x86.avx512.pternlog.d.128
return _mm_ternarylogic_epi32(__A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4si(
+ _mm_ternarylogic_epi32(
+ ((__m128i)((__v4si){-0x1, 0x0, -0x1, 0x0})),
+ ((__m128i)((__v4si){0xB, 0xB, 0xB, 0xB})),
+ ((__m128i)((__v4si){0xC, 0xC, 0xC, 0xC })),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0xB, 0xC));
+TEST_CONSTEXPR(match_v4si(
+ _mm_ternarylogic_epi32(
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4si(
+ _mm_ternarylogic_epi32(
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0));
__m128i test_mm_mask_ternarylogic_epi32(__m128i __A, __mmask8 __U, __m128i
__B, __m128i __C) {
// CHECK-LABEL: test_mm_mask_ternarylogic_epi32
@@ -8335,6 +8356,30 @@ __m128i test_mm_mask_ternarylogic_epi32(__m128i __A,
__mmask8 __U, __m128i __B,
// CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
return _mm_mask_ternarylogic_epi32(__A, __U, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4si(
+ _mm_mask_ternarylogic_epi32(
+ ((__m128i)((__v4si){-0x1, 0x0, -0x1, 0x0})),
+ (__mmask8)0x03,
+ ((__m128i)((__v4si){0xB, 0xB, 0xB, 0xB})),
+ ((__m128i)((__v4si){0xC, 0xC, 0xC, 0xC })),
+ (unsigned char)0xCA),
+ 0xB, 0xC, -0x1, 0x0));
+TEST_CONSTEXPR(match_v4si(
+ _mm_mask_ternarylogic_epi32(
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0x0C,
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4si(
+ _mm_mask_ternarylogic_epi32(
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0x05,
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9, 0x0, 0x9));
__m128i test_mm_maskz_ternarylogic_epi32(__mmask8 __U, __m128i __A, __m128i
__B, __m128i __C) {
// CHECK-LABEL: test_mm_maskz_ternarylogic_epi32
@@ -8342,12 +8387,57 @@ __m128i test_mm_maskz_ternarylogic_epi32(__mmask8 __U,
__m128i __A, __m128i __B,
// CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32>
zeroinitializer
return _mm_maskz_ternarylogic_epi32(__U, __A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x03,
+ ((__m128i)((__v4si){-0x1, 0x0, -0x1, 0x0})),
+ ((__m128i)((__v4si){0xB, 0xB, 0xB, 0xB})),
+ ((__m128i)((__v4si){0xC, 0xC, 0xC, 0xC })),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0x0, 0x0));
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x0C,
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x05,
+ ((__m128i)((__v4si){0x9, 0x9, 0x9, 0x9})),
+ ((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
+ ((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0));
__m256i test_mm256_ternarylogic_epi32(__m256i __A, __m256i __B, __m256i __C) {
// CHECK-LABEL: test_mm256_ternarylogic_epi32
// CHECK: @llvm.x86.avx512.pternlog.d.256
return _mm256_ternarylogic_epi32(__A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v8si(
+ _mm256_ternarylogic_epi32(
+ ((__m256i)((__v8si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m256i)((__v8si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v8si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_ternarylogic_epi32(
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_ternarylogic_epi32(
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0));
__m256i test_mm256_mask_ternarylogic_epi32(__m256i __A, __mmask8 __U, __m256i
__B, __m256i __C) {
// CHECK-LABEL: test_mm256_mask_ternarylogic_epi32
@@ -8355,6 +8445,30 @@ __m256i test_mm256_mask_ternarylogic_epi32(__m256i __A,
__mmask8 __U, __m256i __
// CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}}
return _mm256_mask_ternarylogic_epi32(__A, __U, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v8si(
+ _mm256_mask_ternarylogic_epi32(
+ ((__m256i)((__v8si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ (__mmask8)0x33,
+ ((__m256i)((__v8si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v8si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_mask_ternarylogic_epi32(
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0xCC,
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_mask_ternarylogic_epi32(
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0x55,
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9));
__m256i test_mm256_maskz_ternarylogic_epi32(__mmask8 __U, __m256i __A, __m256i
__B, __m256i __C) {
// CHECK-LABEL: test_mm256_maskz_ternarylogic_epi32
@@ -8362,12 +8476,57 @@ __m256i test_mm256_maskz_ternarylogic_epi32(__mmask8
__U, __m256i __A, __m256i _
// CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32>
zeroinitializer
return _mm256_maskz_ternarylogic_epi32(__U, __A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v8si(
+ _mm256_maskz_ternarylogic_epi32(
+ (__mmask8)0x33,
+ ((__m256i)((__v8si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
+ ((__m256i)((__v8si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v8si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_maskz_ternarylogic_epi32(
+ (__mmask8)0xCC,
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF));
+TEST_CONSTEXPR(match_v8si(
+ _mm256_maskz_ternarylogic_epi32(
+ (__mmask8)0x55,
+ ((__m256i)((__v8si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v8si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v8si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0));
__m128i test_mm_ternarylogic_epi64(__m128i __A, __m128i __B, __m128i __C) {
// CHECK-LABEL: test_mm_ternarylogic_epi64
// CHECK: @llvm.x86.avx512.pternlog.q.128
return _mm_ternarylogic_epi64(__A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v2di(
+ _mm_ternarylogic_epi64(
+ ((__m128i)((__v2di){-0x1, 0x0})),
+ ((__m128i)((__v2di){0xB, 0xB})),
+ ((__m128i)((__v2di){0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC));
+TEST_CONSTEXPR(match_v2di(
+ _mm_ternarylogic_epi64(
+ ((__m128i)((__v2di){0x9, 0x9})),
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF));
+TEST_CONSTEXPR(match_v2di(
+ _mm_ternarylogic_epi64(
+ ((__m128i)((__v2di){0x9, 0x9})),
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0));
__m128i test_mm_mask_ternarylogic_epi64(__m128i __A, __mmask8 __U, __m128i
__B, __m128i __C) {
// CHECK-LABEL: test_mm_mask_ternarylogic_epi64
@@ -8375,6 +8534,30 @@ __m128i test_mm_mask_ternarylogic_epi64(__m128i __A,
__mmask8 __U, __m128i __B,
// CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}
return _mm_mask_ternarylogic_epi64(__A, __U, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v2di(
+ _mm_mask_ternarylogic_epi64(
+ ((__m128i)((__v2di){-0x1, 0x0})),
+ (__mmask8)0x33,
+ ((__m128i)((__v2di){0xB, 0xB})),
+ ((__m128i)((__v2di){0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC));
+TEST_CONSTEXPR(match_v2di(
+ _mm_mask_ternarylogic_epi64(
+ ((__m128i)((__v2di){0x9, 0x9})),
+ (__mmask8)0xCC,
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9));
+TEST_CONSTEXPR(match_v2di(
+ _mm_mask_ternarylogic_epi64(
+ ((__m128i)((__v2di){0x9, 0x9})),
+ (__mmask8)0x55,
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9));
__m128i test_mm_maskz_ternarylogic_epi64(__mmask8 __U, __m128i __A, __m128i
__B, __m128i __C) {
// CHECK-LABEL: test_mm_maskz_ternarylogic_epi64
@@ -8382,12 +8565,57 @@ __m128i test_mm_maskz_ternarylogic_epi64(__mmask8 __U,
__m128i __A, __m128i __B,
// CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64>
zeroinitializer
return _mm_maskz_ternarylogic_epi64(__U, __A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v2di(
+ _mm_maskz_ternarylogic_epi64(
+ (__mmask8)0x03,
+ ((__m128i)((__v2di){-0x1, 0x0})),
+ ((__m128i)((__v2di){0xB, 0xB})),
+ ((__m128i)((__v2di){0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC));
+TEST_CONSTEXPR(match_v2di(
+ _mm_maskz_ternarylogic_epi64(
+ (__mmask8)0x0C,
+ ((__m128i)((__v2di){0x9, 0x9})),
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0));
+TEST_CONSTEXPR(match_v2di(
+ _mm_maskz_ternarylogic_epi64(
+ (__mmask8)0x05,
+ ((__m128i)((__v2di){0x9, 0x9})),
+ ((__m128i)((__v2di){0x4, 0x4})),
+ ((__m128i)((__v2di){0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0));
__m256i test_mm256_ternarylogic_epi64(__m256i __A, __m256i __B, __m256i __C) {
// CHECK-LABEL: test_mm256_ternarylogic_epi64
// CHECK: @llvm.x86.avx512.pternlog.q.256
return _mm256_ternarylogic_epi64(__A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4di(
+ _mm256_ternarylogic_epi64(
+ ((__m256i)((__v4di){-0x1, 0x0, -0x1, 0x0})),
+ ((__m256i)((__v4di){0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v4di){0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0xB, 0xC));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_ternarylogic_epi64(
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0xF, 0xF, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_ternarylogic_epi64(
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0));
__m256i test_mm256_mask_ternarylogic_epi64(__m256i __A, __mmask8 __U, __m256i
__B, __m256i __C) {
// CHECK-LABEL: test_mm256_mask_ternarylogic_epi64
@@ -8395,6 +8623,30 @@ __m256i test_mm256_mask_ternarylogic_epi64(__m256i __A,
__mmask8 __U, __m256i __
// CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}
return _mm256_mask_ternarylogic_epi64(__A, __U, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4di(
+ _mm256_mask_ternarylogic_epi64(
+ ((__m256i)((__v4di){-0x1, 0x0, -0x1, 0x0})),
+ (__mmask8)0x33,
+ ((__m256i)((__v4di){0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v4di){0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, -0x1, 0x0));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_mask_ternarylogic_epi64(
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0xCC,
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x9, 0x9, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_mask_ternarylogic_epi64(
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ (__mmask8)0x55,
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x9, 0x0, 0x9));
__m256i test_mm256_maskz_ternarylogic_epi64(__mmask8 __U, __m256i __A, __m256i
__B, __m256i __C) {
// CHECK-LABEL: test_mm256_maskz_ternarylogic_epi64
@@ -8402,6 +8654,31 @@ __m256i test_mm256_maskz_ternarylogic_epi64(__mmask8
__U, __m256i __A, __m256i _
// CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64>
zeroinitializer
return _mm256_maskz_ternarylogic_epi64(__U, __A, __B, __C, 4);
}
+TEST_CONSTEXPR(match_v4di(
+ _mm256_maskz_ternarylogic_epi64(
+ (__mmask8)0x33,
+ ((__m256i)((__v4di){-0x1, 0x0, -0x1, 0x0})),
+ ((__m256i)((__v4di){0xB, 0xB, 0xB, 0xB})),
+ ((__m256i)((__v4di){0xC, 0xC, 0xC, 0xC})),
+ (unsigned char)0xCA),
+ 0xB, 0xC, 0x0, 0x0));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_maskz_ternarylogic_epi64(
+ (__mmask8)0xCC,
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0xFE),
+ 0x0, 0x0, 0xF, 0xF));
+TEST_CONSTEXPR(match_v4di(
+ _mm256_maskz_ternarylogic_epi64(
+ (__mmask8)0x55,
+ ((__m256i)((__v4di){0x9, 0x9, 0x9, 0x9})),
+ ((__m256i)((__v4di){0x4, 0x4, 0x4, 0x4})),
+ ((__m256i)((__v4di){0x2, 0x2, 0x2, 0x2})),
+ (unsigned char)0x80),
+ 0x0, 0x0, 0x0, 0x0));
+
__m256 test_mm256_shuffle_f32x4(__m256 __A, __m256 __B) {
// CHECK-LABEL: test_mm256_shuffle_f32x4
// CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32>
<i32 4, i32 5, i32 6, i32 7, i32 12, i32 13, i32 14, i32 15>
>From b943ad28ee5054777c7431237dd4561dc6bdfe33 Mon Sep 17 00:00:00 2001
From: Shawn <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:05:54 -0700
Subject: [PATCH 02/14] Apply suggestion from @tbaederr
Co-authored-by: Timm Baeder <tbae...@redhat.com>
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 00466aa7ae663..cd376bee4f288 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2971,7 +2971,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const Pointer &Dst = S.Stk.peek<Pointer>();
- for (unsigned I = 0; I < DstLen; ++I) {
+ for (unsigned I = 0; I != DstLen; ++I) {
APSInt a, b, c;
INT_TYPE_SWITCH(DstElemT, {
a = A.elem<T>(I).toAPSInt();
>From d2eb52bee393f449af35e49b7164013e896245c9 Mon Sep 17 00:00:00 2001
From: Shawn <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:06:30 -0700
Subject: [PATCH 03/14] Apply suggestion from @tbaederr
Co-authored-by: Timm Baeder <tbae...@redhat.com>
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index cd376bee4f288..3c39c683805d2 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2984,7 +2984,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
bool DstUnsigned = a.isUnsigned();
if (U[I]) {
- for (unsigned Bit = 0; Bit < BitWidth; ++Bit) {
+ for (unsigned Bit = 0; Bit != BitWidth; ++Bit) {
unsigned Idx = (a[Bit] << 2) | (b[Bit] << 1) | (c[Bit]);
R.setBitVal(Bit, Imm[Idx]);
}
>From 26b535489fbc6a01391483274390359de08be6f5 Mon Sep 17 00:00:00 2001
From: Shawn <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:09:14 -0700
Subject: [PATCH 04/14] Apply suggestion from @tbaederr
Co-authored-by: Timm Baeder <tbae...@redhat.com>
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 3c39c683805d2..8992c23a4fd6c 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -3659,7 +3659,7 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const
CallExpr *Call,
case X86::BI__builtin_ia32_pternlogq128_mask:
case X86::BI__builtin_ia32_pternlogq256_mask:
case X86::BI__builtin_ia32_pternlogq512_mask:
- return interp__builtin_pternlog(S, OpPC, Call, false);
+ return interp__builtin_pternlog(S, OpPC, Call, /*MaskZ=*/false);
case X86::BI__builtin_ia32_pternlogd128_maskz:
case X86::BI__builtin_ia32_pternlogd256_maskz:
case X86::BI__builtin_ia32_pternlogd512_maskz:
>From 2f1a3895d594e843a619c1fbfa44aff592d0dac0 Mon Sep 17 00:00:00 2001
From: Shawn <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:09:42 -0700
Subject: [PATCH 05/14] Apply suggestion from @tbaederr
Co-authored-by: Timm Baeder <tbae...@redhat.com>
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 8992c23a4fd6c..ccf0a47558202 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -3666,7 +3666,7 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const
CallExpr *Call,
case X86::BI__builtin_ia32_pternlogq128_maskz:
case X86::BI__builtin_ia32_pternlogq256_maskz:
case X86::BI__builtin_ia32_pternlogq512_maskz:
- return interp__builtin_pternlog(S, OpPC, Call, true);
+ return interp__builtin_pternlog(S, OpPC, Call, /*MaskZ=*/true);
case Builtin::BI__builtin_elementwise_fshl:
return interp__builtin_elementwise_triop(S, OpPC, Call,
llvm::APIntOps::fshl);
>From f20253020ba0e4174a0812f040af2c6476335050 Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:47:53 -0700
Subject: [PATCH 06/14] Apply feedback: Use PascalCase
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 26 ++++++++++++------------
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index ccf0a47558202..88a17df1b5e4a 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2972,32 +2972,32 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const Pointer &Dst = S.Stk.peek<Pointer>();
for (unsigned I = 0; I != DstLen; ++I) {
- APSInt a, b, c;
+ APSInt ALane, BLane, CLane;
INT_TYPE_SWITCH(DstElemT, {
- a = A.elem<T>(I).toAPSInt();
- b = B.elem<T>(I).toAPSInt();
- c = C.elem<T>(I).toAPSInt();
+ ALane = A.elem<T>(I).toAPSInt();
+ BLane = B.elem<T>(I).toAPSInt();
+ CLane = C.elem<T>(I).toAPSInt();
});
- unsigned BitWidth = a.getBitWidth();
- APInt R(BitWidth, 0);
- bool DstUnsigned = a.isUnsigned();
+ unsigned BitWidth = ALane.getBitWidth();
+ APInt RLane(BitWidth, 0);
+ bool DstUnsigned = ALane.isUnsigned();
if (U[I]) {
for (unsigned Bit = 0; Bit != BitWidth; ++Bit) {
- unsigned Idx = (a[Bit] << 2) | (b[Bit] << 1) | (c[Bit]);
- R.setBitVal(Bit, Imm[Idx]);
+ unsigned Idx = (ALane[Bit] << 2) | (BLane[Bit] << 1) | (CLane[Bit]);
+ RLane.setBitVal(Bit, Imm[Idx]);
}
INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
- Dst.elem<T>(I) = static_cast<T>(APSInt(R, DstUnsigned));
+ Dst.elem<T>(I) = static_cast<T>(APSInt(RLane, DstUnsigned));
});
} else if (MaskZ) {
- INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
- Dst.elem<T>(I) = static_cast<T>(APSInt(R, DstUnsigned));
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT, { /* Zeroes lane */
+ Dst.elem<T>(I) = static_cast<T>(APSInt(RLane, DstUnsigned));
});
} else {
INT_TYPE_SWITCH_NO_BOOL(DstElemT,
- { Dst.elem<T>(I) = static_cast<T>(a); });
+ { Dst.elem<T>(I) = static_cast<T>(ALane); });
}
}
Dst.initializeAllElements();
>From 5e76351fc76d0ae932ba8dbf1840416a0a96459d Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Tue, 16 Sep 2025 00:53:17 -0700
Subject: [PATCH 07/14] Clang-format
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 88a17df1b5e4a..cbfeed1c5e5d5 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2992,7 +2992,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
Dst.elem<T>(I) = static_cast<T>(APSInt(RLane, DstUnsigned));
});
} else if (MaskZ) {
- INT_TYPE_SWITCH_NO_BOOL(DstElemT, { /* Zeroes lane */
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT, { // Zeroes lane
Dst.elem<T>(I) = static_cast<T>(APSInt(RLane, DstUnsigned));
});
} else {
>From 9a227c02ca6eb0ac052f91b8d14fbad3143e3051 Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Wed, 17 Sep 2025 23:26:28 -0700
Subject: [PATCH 08/14] Apply feedback: Rebase and refactor lines
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index cbfeed1c5e5d5..be2eea5feb78e 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2961,10 +2961,10 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
unsigned DstLen = VecT->getNumElements();
- PrimType DstElemT = *S.getContext().classify(VecT->getElementType());
+ const PrimType &DstElemT = *S.getContext().classify(VecT->getElementType());
- APSInt U = popToAPSInt(S.Stk, *S.getContext().classify(Call->getArg(4)));
- APSInt Imm = popToAPSInt(S.Stk, *S.getContext().classify(Call->getArg(3)));
+ APSInt U = popToAPSInt(S, Call->getArg(4));
+ APSInt Imm = popToAPSInt(S, Call->getArg(3));
const Pointer &C = S.Stk.pop<Pointer>();
const Pointer &B = S.Stk.pop<Pointer>();
const Pointer &A = S.Stk.pop<Pointer>();
>From 9cf5aef34c85ae148d76a04a8423790ee89c0f99 Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Fri, 19 Sep 2025 14:41:05 -0700
Subject: [PATCH 09/14] Apply feedback: Rebase with main
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 1 +
1 file changed, 1 insertion(+)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index be2eea5feb78e..e23d4798c6706 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -3001,6 +3001,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
}
}
Dst.initializeAllElements();
+ return true;
}
bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
>From decd3b81ef2fa011f7cfbb3a8ac8cb4d5a3e27bc Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Fri, 19 Sep 2025 14:47:39 -0700
Subject: [PATCH 10/14] Apply feedback: Rebase with main
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 1 -
1 file changed, 1 deletion(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index e23d4798c6706..5b84783b0ff40 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2978,7 +2978,6 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
BLane = B.elem<T>(I).toAPSInt();
CLane = C.elem<T>(I).toAPSInt();
});
-
unsigned BitWidth = ALane.getBitWidth();
APInt RLane(BitWidth, 0);
bool DstUnsigned = ALane.isUnsigned();
>From 3843b3a6258895953c00b484c9e98823c21a61ef Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Mon, 22 Sep 2025 21:16:42 -0700
Subject: [PATCH 11/14] Upload failing testcase for now
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 10 ++-
clang/test/CodeGen/X86/avx512f-builtins.c | 89 +++++++++++++----------
2 files changed, 58 insertions(+), 41 deletions(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 5b84783b0ff40..f2af4a90c1f9a 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2960,7 +2960,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
assert(Call->getNumArgs() == 5);
const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
- unsigned DstLen = VecT->getNumElements();
+ const unsigned DstLen = VecT->getNumElements();
const PrimType &DstElemT = *S.getContext().classify(VecT->getElementType());
APSInt U = popToAPSInt(S, Call->getArg(4));
@@ -2973,7 +2973,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
for (unsigned I = 0; I != DstLen; ++I) {
APSInt ALane, BLane, CLane;
- INT_TYPE_SWITCH(DstElemT, {
+ INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
ALane = A.elem<T>(I).toAPSInt();
BLane = B.elem<T>(I).toAPSInt();
CLane = C.elem<T>(I).toAPSInt();
@@ -2984,7 +2984,11 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
if (U[I]) {
for (unsigned Bit = 0; Bit != BitWidth; ++Bit) {
- unsigned Idx = (ALane[Bit] << 2) | (BLane[Bit] << 1) | (CLane[Bit]);
+ unsigned ABit = ALane[Bit];
+ unsigned BBit = BLane[Bit];
+ unsigned CBit = CLane[Bit];
+
+ unsigned Idx = (ABit << 2) | (BBit << 1) | (CBit);
RLane.setBitVal(Bit, Imm[Idx]);
}
INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
diff --git a/clang/test/CodeGen/X86/avx512f-builtins.c
b/clang/test/CodeGen/X86/avx512f-builtins.c
index 9fa450c68e503..223551233e88f 100644
--- a/clang/test/CodeGen/X86/avx512f-builtins.c
+++ b/clang/test/CodeGen/X86/avx512f-builtins.c
@@ -6251,21 +6251,21 @@ TEST_CONSTEXPR(match_v16si(
((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
- (unsigned char)0xCA),
+ (unsigned char)0xCA), // A ? B : C
0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB,
0xC));
TEST_CONSTEXPR(match_v16si(
_mm512_ternarylogic_epi32(
((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0xFE),
+ (unsigned char)0xFE), // A | B | C
0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF,
0xF));
TEST_CONSTEXPR(match_v16si(
_mm512_ternarylogic_epi32(
((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0x80),
+ (unsigned char)0x80), // A & B & C
0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0));
__m512i test_mm512_mask_ternarylogic_epi32(__m512i __A, __mmask16 __U, __m512i
__B, __m512i __C) {
@@ -6276,28 +6276,37 @@ __m512i test_mm512_mask_ternarylogic_epi32(__m512i __A,
__mmask16 __U, __m512i _
}
TEST_CONSTEXPR(match_v16si(
_mm512_mask_ternarylogic_epi32(
- ((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
- (__mmask16)0x3333,
- ((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
- ((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
- (unsigned char)0xCA),
- 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0, 0xB, 0xC, -0x1, 0x0, 0xB, 0xC,
-0x1, 0x0));
+ ((__m512i)((__v16si){0x1, 0x0, 0x2, 0x0, 0x3, 0x0, 0x4, 0x0,
+ 0x5, 0x0, 0x6, 0x0, 0x7, 0x0, 0x8, 0x0})),
+ (__mmask16)0xA55A,
+ ((__m512i)((__v16si){0xA, 0xB, 0xC, 0xD, 0xE, 0xF, 0x10, 0x11,
+ 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19})),
+ ((__m512i)((__v16si){0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8,
+ 0x9, 0xA, 0xB, 0xC, 0xD, 0xE, 0xF, 0x10})),
+ (unsigned char)0xCA), // A ? B : C
+ 0x1, 0x2, 0x2, 0x4, 0x6, 0x0, 0x3, 0x0, 0x8, 0x0, 0xD, 0x0, 0x7, 0xE, 0x8,
0x10));
TEST_CONSTEXPR(match_v16si(
_mm512_mask_ternarylogic_epi32(
- ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
- (__mmask16)0xCCCC,
- ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
- ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0xFE),
- 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF, 0xF, 0x9, 0x9, 0xF,
0xF));
+ ((__m512i)((__v16si){0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7,
+ 0x8, 0x9, 0xA, 0xB, 0xC, 0xD, 0xE, 0xF})),
+ (__mmask16)0x0F0F,
+ ((__m512i)((__v16si){0x1, 0x2, 0x4, 0x8, 0x1, 0x2, 0x4, 0x8,
+ 0x1, 0x2, 0x4, 0x8, 0x1, 0x2, 0x4, 0x8})),
+ ((__m512i)((__v16si){0x10, 0x20, 0x40, 0x80, 0x10, 0x20, 0x40, 0x80,
+ 0x10, 0x20, 0x40, 0x80, 0x10, 0x20, 0x40, 0x80})),
+ (unsigned char)0xFE), // A | B | C
+ 0x11, 0x23, 0x46, 0x8B, 0x4, 0x5, 0x6, 0x7, 0x19, 0x2B, 0x4E, 0x8B, 0xC,
0xD, 0xE, 0xF));
TEST_CONSTEXPR(match_v16si(
_mm512_mask_ternarylogic_epi32(
- ((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
- (__mmask16)0x5555,
- ((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
- ((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0x80),
- 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0, 0x9, 0x0,
0x9));
+ ((__m512i)((__v16si){0xF, 0x7, 0x3, 0x1, 0xF, 0x7, 0x3, 0x1,
+ 0xFF, 0xF, 0xF0, 0xAA, 0x55, 0xCC, 0x33, 0xFF})),
+ (__mmask16)0xAAAA,
+ ((__m512i)((__v16si){0xE, 0x7, 0x2, 0x1, 0xF, 0x0, 0x3, 0x0,
+ 0xF, 0xF0, 0xFF, 0x55, 0x55, 0x33, 0x33, 0xF})),
+ ((__m512i)((__v16si){0xD, 0x7, 0x0, 0x1, 0xF, 0x7, 0x0, 0x1,
+ 0xF0, 0xF, 0xF, 0xFF, 0xF, 0xCC, 0x33, 0xF0})),
+ (unsigned char)0x80), // A & B & C
+ 0xF, 0x7, 0x3, 0x1, 0xF, 0x0, 0x3, 0x0, 0xFF, 0x0, 0xF0, 0x0, 0x55, 0x0,
0x33, 0x0));
__m512i test_mm512_maskz_ternarylogic_epi32(__mmask16 __U, __m512i __A,
__m512i __B, __m512i __C) {
// CHECK-LABEL: test_mm512_maskz_ternarylogic_epi32
@@ -6311,7 +6320,7 @@ TEST_CONSTEXPR(match_v16si(
((__m512i)((__v16si){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1,
0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
((__m512i)((__v16si){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB,
0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
((__m512i)((__v16si){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC,
0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
- (unsigned char)0xCA),
+ (unsigned char)0xCA), // A ? B : C
0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0, 0x0, 0xB, 0xC, 0x0,
0x0));
TEST_CONSTEXPR(match_v16si(
_mm512_maskz_ternarylogic_epi32(
@@ -6319,7 +6328,7 @@ TEST_CONSTEXPR(match_v16si(
((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0xFE),
+ (unsigned char)0xFE), // A | B | C
0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF, 0xF, 0x0, 0x0, 0xF,
0xF));
TEST_CONSTEXPR(match_v16si(
_mm512_maskz_ternarylogic_epi32(
@@ -6327,7 +6336,7 @@ TEST_CONSTEXPR(match_v16si(
((__m512i)((__v16si){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9,
0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
((__m512i)((__v16si){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4,
0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
((__m512i)((__v16si){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2,
0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0x80),
+ (unsigned char)0x80), // A & B & C
0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0));
__m512i test_mm512_ternarylogic_epi64(__m512i __A, __m512i __B, __m512i __C) {
@@ -6337,25 +6346,29 @@ __m512i test_mm512_ternarylogic_epi64(__m512i __A,
__m512i __B, __m512i __C) {
}
TEST_CONSTEXPR(match_v8di(
_mm512_ternarylogic_epi64(
+ ((__m512i)((__v8di){0x1111, 0x2222, 0x3333, 0x4444, 0x5555, 0x6666,
0x7777, 0x8888})),
+ ((__m512i)((__v8di){0xAAAA, 0xBBBB, 0xCCCC, 0xDDDD, 0xEEEE, 0xFFFF,
0x1111, 0x2222})),
((__m512i)((__v8di){-0x1, 0x0, -0x1, 0x0, -0x1, 0x0, -0x1, 0x0})),
- ((__m512i)((__v8di){0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB, 0xB})),
- ((__m512i)((__v8di){0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC, 0xC})),
- (unsigned char)0xCA),
- 0xB, 0xC, 0xB, 0xC, 0xB, 0xC, 0xB, 0xC));
+ (unsigned char)0xD8), // C ? B : A
+ 0xAAAA, 0x2222, 0xCCCC, 0x4444, 0xEEEE, 0x6666, 0x1111, 0x8888));
TEST_CONSTEXPR(match_v8di(
_mm512_ternarylogic_epi64(
- ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
- ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
- ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0xFE),
- 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF, 0xF));
+ ((__m512i)((__v8di){-0x1, 0x0, -0x1, 0x0, 0xF0F0, 0xFF,
-0x5555555555555556, 0x5555555555555555})),
+ ((__m512i)((__v8di){0x1234, 0xFFFF, 0xFF, 0xF0F, 0x3333, 0xFF00,
-0x5555555555555556, -0x0F0F0F0F0F0F0F10})),
+ ((__m512i)((__v8di){0xFFFF, 0x1234, 0xF0F, 0xFF00, 0xF0F0, 0x3333,
0x5555555555555555, 0x0F0F0F0F0F0F0F0})),
+ (unsigned char)0x8F), // ~A | (B & C)
+ 0x1234, -0x1, 0xF, -0x1, -0xC0C1, -0x100, 0x5555555555555555,
-0x5505050505050506));
TEST_CONSTEXPR(match_v8di(
_mm512_ternarylogic_epi64(
- ((__m512i)((__v8di){0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9, 0x9})),
- ((__m512i)((__v8di){0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4, 0x4})),
- ((__m512i)((__v8di){0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0x80),
- 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0));
+ ((__m512i)((__v8di){0x7FFFFFFFFFFFFFFF, 0x0, 0x00FF00FF00FF00FF,
0x0F0F0F0F0F0F0F0F,
+ 0x123456789ABCDEF0, 0x3333333333333333,
0x5555555555555555, 0x0123456789ABCDEF})),
+ ((__m512i)((__v8di){0x1111111111111111, 0x2222222222222222, 0xFFFFFFFF,
-0x100000000,
+ 0x0, -0x3333333333333334, -0x0F0F0F0F0F0F0F10,
-0x123456789ABCDF0})),
+ ((__m512i)((__v8di){0x2222222222222222, 0x1111111111111111,
-0x1000000000000, 0xFFFFFFFF,
+ -0x1, 0x0, 0x0F0F0F0F0F0F0F0F, 0x0})),
+ (unsigned char)0xE0), // A & (B | C)
+ 0x3333333333333333, 0x0, 0x00FF000000FF00FF, 0x0F0F0F0F0F0F0F0F,
+ 0x123456789ABCDEF0, 0x0, 0x5555555555555555, 0x0));
__m512i test_mm512_mask_ternarylogic_epi64(__m512i __A, __mmask8 __U, __m512i
__B, __m512i __C) {
// CHECK-LABEL: test_mm512_mask_ternarylogic_epi64
>From 1d8e4bff57c99b8f7de8503d583501528c5fe9f0 Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Wed, 24 Sep 2025 08:14:40 -0700
Subject: [PATCH 12/14] Save debug changes
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index f2af4a90c1f9a..bc9e5e0ab839e 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -2962,9 +2962,10 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
const unsigned DstLen = VecT->getNumElements();
const PrimType &DstElemT = *S.getContext().classify(VecT->getElementType());
+ const bool DstUnsigned = VecT->isUnsignedIntegerOrEnumerationType();
- APSInt U = popToAPSInt(S, Call->getArg(4));
- APSInt Imm = popToAPSInt(S, Call->getArg(3));
+ APInt U = popToAPSInt(S, Call->getArg(4));
+ APInt Imm = popToAPSInt(S, Call->getArg(3));
const Pointer &C = S.Stk.pop<Pointer>();
const Pointer &B = S.Stk.pop<Pointer>();
const Pointer &A = S.Stk.pop<Pointer>();
@@ -2972,7 +2973,9 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const Pointer &Dst = S.Stk.peek<Pointer>();
for (unsigned I = 0; I != DstLen; ++I) {
- APSInt ALane, BLane, CLane;
+ APInt ALane;
+ APInt BLane;
+ APInt CLane;
INT_TYPE_SWITCH_NO_BOOL(DstElemT, {
ALane = A.elem<T>(I).toAPSInt();
BLane = B.elem<T>(I).toAPSInt();
@@ -2980,7 +2983,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
});
unsigned BitWidth = ALane.getBitWidth();
APInt RLane(BitWidth, 0);
- bool DstUnsigned = ALane.isUnsigned();
+ // bool DstUnsigned = ALane.isUnsigned();
if (U[I]) {
for (unsigned Bit = 0; Bit != BitWidth; ++Bit) {
@@ -3000,7 +3003,7 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
});
} else {
INT_TYPE_SWITCH_NO_BOOL(DstElemT,
- { Dst.elem<T>(I) = static_cast<T>(ALane); });
+ { Dst.elem<T>(I) = static_cast<T>(APSInt(ALane,
DstUnsigned)); });
}
}
Dst.initializeAllElements();
>From b0ae797a6e2aa3ed8af8a0852e6dab42ef1f499b Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Thu, 25 Sep 2025 18:52:33 -0700
Subject: [PATCH 13/14] Save debug changes
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index bc9e5e0ab839e..18001737a783d 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -20,6 +20,10 @@
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/SipHash.h"
+#include <iostream>
+#include "llvm/Support/Debug.h"
+#define DEBUG_TYPE "interp-builtin"
+
namespace clang {
namespace interp {
@@ -2959,6 +2963,9 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const CallExpr *Call, bool MaskZ) {
assert(Call->getNumArgs() == 5);
+ // LLVM_DEBUG(llvm::dbgs() << "Debug\n");
+ std::cout << "Debug here\n";
+
const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
const unsigned DstLen = VecT->getNumElements();
const PrimType &DstElemT = *S.getContext().classify(VecT->getElementType());
>From 347bfb5685b8063fd97199c262c9c09a3f3f707a Mon Sep 17 00:00:00 2001
From: kimsh02 <kimshaw...@icloud.com>
Date: Sat, 27 Sep 2025 21:30:33 -0700
Subject: [PATCH 14/14] Save debug changes
---
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 23 +++++++-------
clang/test/CodeGen/X86/avx512f-builtins.c | 5 +--
clang/test/CodeGen/X86/avx512vl-builtins.c | 36 +++++++++++++++++++---
3 files changed, 44 insertions(+), 20 deletions(-)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 18001737a783d..b06016f5f54c3 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -20,10 +20,6 @@
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/SipHash.h"
-#include <iostream>
-#include "llvm/Support/Debug.h"
-#define DEBUG_TYPE "interp-builtin"
-
namespace clang {
namespace interp {
@@ -2963,16 +2959,13 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
const CallExpr *Call, bool MaskZ) {
assert(Call->getNumArgs() == 5);
- // LLVM_DEBUG(llvm::dbgs() << "Debug\n");
- std::cout << "Debug here\n";
-
const VectorType *VecT = Call->getArg(0)->getType()->castAs<VectorType>();
const unsigned DstLen = VecT->getNumElements();
const PrimType &DstElemT = *S.getContext().classify(VecT->getElementType());
- const bool DstUnsigned = VecT->isUnsignedIntegerOrEnumerationType();
+ const bool DstUnsigned =
VecT->getElementType()->isUnsignedIntegerOrEnumerationType();
- APInt U = popToAPSInt(S, Call->getArg(4));
- APInt Imm = popToAPSInt(S, Call->getArg(3));
+ const APInt U = popToAPSInt(S, Call->getArg(4));
+ const APInt Imm = popToAPSInt(S, Call->getArg(3));
const Pointer &C = S.Stk.pop<Pointer>();
const Pointer &B = S.Stk.pop<Pointer>();
const Pointer &A = S.Stk.pop<Pointer>();
@@ -2988,9 +2981,17 @@ static bool interp__builtin_pternlog(InterpState &S,
CodePtr OpPC,
BLane = B.elem<T>(I).toAPSInt();
CLane = C.elem<T>(I).toAPSInt();
});
- unsigned BitWidth = ALane.getBitWidth();
+ const unsigned BitWidth = ALane.getBitWidth();
APInt RLane(BitWidth, 0);
// bool DstUnsigned = ALane.isUnsigned();
+
+ #define DEBUG_TYPE "ptern"
+ LLVM_DEBUG({
+ ALane.print(llvm::dbgs(), false); llvm::dbgs() << "\n";
+ BLane.print(llvm::dbgs(), false); llvm::dbgs() << "\n";
+ CLane.print(llvm::dbgs(), false); llvm::dbgs() << "\n";
+ RLane.print(llvm::dbgs(), false); llvm::dbgs() << "\n";
+ });
if (U[I]) {
for (unsigned Bit = 0; Bit != BitWidth; ++Bit) {
diff --git a/clang/test/CodeGen/X86/avx512f-builtins.c
b/clang/test/CodeGen/X86/avx512f-builtins.c
index 223551233e88f..41be6de64ed1a 100644
--- a/clang/test/CodeGen/X86/avx512f-builtins.c
+++ b/clang/test/CodeGen/X86/avx512f-builtins.c
@@ -3,10 +3,7 @@
// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall
-Werror -Wsign-conversion | FileCheck %s
// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -fms-extensions
-fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc
-target-feature +avx512f -emit-llvm -o - -Wall -Werror -Wsign-conversion |
FileCheck %s
-// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall
-Werror -Wsign-conversion -fexperimental-new-constant-interpreter | FileCheck %s
-// RUN: %clang_cc1 -x c -flax-vector-conversions=none -fms-extensions
-fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc
-target-feature +avx512f -emit-llvm -o - -Wall -Werror -Wsign-conversion
-fexperimental-new-constant-interpreter | FileCheck %s
-// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall
-Werror -Wsign-conversion -fexperimental-new-constant-interpreter | FileCheck %s
-// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -fms-extensions
-fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc
-target-feature +avx512f -emit-llvm -o - -Wall -Werror -Wsign-conversion
-fexperimental-new-constant-interpreter | FileCheck %s
+
#include <immintrin.h>
#include "builtin_test_helpers.h"
diff --git a/clang/test/CodeGen/X86/avx512vl-builtins.c
b/clang/test/CodeGen/X86/avx512vl-builtins.c
index d29045af6e303..8501ab7b306b7 100644
--- a/clang/test/CodeGen/X86/avx512vl-builtins.c
+++ b/clang/test/CodeGen/X86/avx512vl-builtins.c
@@ -1,8 +1,7 @@
// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl
-emit-llvm -o - -Wall -Werror -Wsign-conversion | FileCheck %s
// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl
-emit-llvm -o - -Wall -Werror -Wsign-conversion | FileCheck %s
-// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl
-emit-llvm -o - -Wall -Werror -Wsign-conversion
-fexperimental-new-constant-interpreter | FileCheck %s
-// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s
-triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl
-emit-llvm -o - -Wall -Werror -Wsign-conversion
-fexperimental-new-constant-interpreter | FileCheck %s
+
#include <immintrin.h>
#include "builtin_test_helpers.h"
@@ -8362,7 +8361,7 @@ TEST_CONSTEXPR(match_v4si(
(__mmask8)0x03,
((__m128i)((__v4si){0xB, 0xB, 0xB, 0xB})),
((__m128i)((__v4si){0xC, 0xC, 0xC, 0xC })),
- (unsigned char)0xCA),
+ (unsigned char)0xCA), // A ? B : C
0xB, 0xC, -0x1, 0x0));
TEST_CONSTEXPR(match_v4si(
_mm_mask_ternarylogic_epi32(
@@ -8370,7 +8369,7 @@ TEST_CONSTEXPR(match_v4si(
(__mmask8)0x0C,
((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0xFE),
+ (unsigned char)0xFE), // A | B | C
0x9, 0x9, 0xF, 0xF));
TEST_CONSTEXPR(match_v4si(
_mm_mask_ternarylogic_epi32(
@@ -8378,7 +8377,7 @@ TEST_CONSTEXPR(match_v4si(
(__mmask8)0x05,
((__m128i)((__v4si){0x4, 0x4, 0x4, 0x4})),
((__m128i)((__v4si){0x2, 0x2, 0x2, 0x2})),
- (unsigned char)0x80),
+ (unsigned char)0x80), // A & B & C
0x0, 0x9, 0x0, 0x9));
__m128i test_mm_maskz_ternarylogic_epi32(__mmask8 __U, __m128i __A, __m128i
__B, __m128i __C) {
@@ -8387,6 +8386,33 @@ __m128i test_mm_maskz_ternarylogic_epi32(__mmask8 __U,
__m128i __A, __m128i __B,
// CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32>
zeroinitializer
return _mm_maskz_ternarylogic_epi32(__U, __A, __B, __C, 4);
}
+// B ? A : C imm = 0xE2 (Idx = (A<<2)|(B<<1)|C per VPTERNLOG)
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x0B,
+ ((__m128i)((__v4si){(int)0xDEADBEEF, 0, (int)0xFFFFFFFF, 0x13579BDF})), //
A
+ ((__m128i)((__v4si){(int)0xFFFFFFFF, 0, (int)0xFFFFFFFF, 0})), //
B
+ ((__m128i)((__v4si){(int)0xCAFEBABE, (int)0xFFFFFFFF, 0, 0x2468ACE0})), //
C
+ (unsigned char)0xE2),
+ (int)0xDEADBEEF, (int)0xFFFFFFFF, 0, 0x2468ACE0));
+ // ~(A & B) | ~(B & C) imm = 0x7F
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x0C,
+ ((__m128i)((__v4si){0, (int)0xFFFFFFFF, (int)0xAAAAAAAA, 0x55555555})), //
A
+ ((__m128i)((__v4si){(int)0xFFFFFFFF, 0, (int)0xFFFFFFFF,
(int)0xFFFFFFFF})), // B
+ ((__m128i)((__v4si){(int)0xF0F0F0F0, 0, 0, (int)0xFFFFFFFF})),
// C
+ (unsigned char)0x7F),
+ 0, 0, (int)0xFFFFFFFF, (int)0xAAAAAAAA));
+ // ~A | ~B | C imm = 0xBF
+TEST_CONSTEXPR(match_v4si(
+ _mm_maskz_ternarylogic_epi32(
+ (__mmask8)0x05,
+ ((__m128i)((__v4si){(int)0xFFFFFFFF, 0, 0x12345678, 0})), //
A
+ ((__m128i)((__v4si){0, 0, 0x0000FFFF, (int)0xFFFFFFFF})), //
B
+ ((__m128i)((__v4si){0, 0, 0x0000000F, 0})), //
C
+ (unsigned char)0xBF),
+ (int)0xFFFFFFFF, 0, (int)0xFFFFA98F, 0));
TEST_CONSTEXPR(match_v4si(
_mm_maskz_ternarylogic_epi32(
(__mmask8)0x03,
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
