https://github.com/jhuber6 updated
https://github.com/llvm/llvm-project/pull/157128
>From 97ee7e3b1cfc06b434b71987b9b0a9fd8b14a236 Mon Sep 17 00:00:00 2001
From: Joseph Huber <hube...@outlook.com>
Date: Fri, 5 Sep 2025 10:42:07 -0500
Subject: [PATCH 1/2] [Clang] Rename elementwise builtins to `clzg` and `ctzg`
Summary:
The added bit counting builtins for vectors used `cttz` and `ctlz`,
which is consistent with the LLVM naming convention. However, these are
clang builtins and implement exactly the `__builtin_ctzg` and
`__builtin_clzg` behavior. It is confusing to people familiar with other
other builtins that these are the only bit counting intrinsics named
differently. This includes the additional operation for the undefined
zero case, which was added as a `clzg` extension.
---
clang/docs/LanguageExtensions.rst | 6 +-
clang/include/clang/Basic/Builtins.td | 4 +-
.../include/clang/Basic/DiagnosticASTKinds.td | 2 +-
clang/lib/AST/ByteCode/InterpBuiltin.cpp | 8 +-
clang/lib/AST/ExprConstant.cpp | 22 +++---
clang/lib/CodeGen/CGBuiltin.cpp | 12 +--
clang/lib/Headers/avx512cdintrin.h | 4 +-
clang/lib/Headers/avx512vlcdintrin.h | 8 +-
clang/lib/Sema/SemaChecking.cpp | 4 +-
.../test/CodeGen/builtins-elementwise-math.c | 40 +++++-----
clang/test/Sema/builtins-elementwise-math.c | 28 +++----
clang/test/Sema/constant-builtins-vector.cpp | 74 +++++++++----------
12 files changed, 106 insertions(+), 106 deletions(-)
diff --git a/clang/docs/LanguageExtensions.rst
b/clang/docs/LanguageExtensions.rst
index ad190eace5b05..a150b1c73bc92 100644
--- a/clang/docs/LanguageExtensions.rst
+++ b/clang/docs/LanguageExtensions.rst
@@ -772,7 +772,7 @@ The elementwise intrinsics
``__builtin_elementwise_popcount``,
``__builtin_elementwise_bitreverse``, ``__builtin_elementwise_add_sat``,
``__builtin_elementwise_sub_sat``, ``__builtin_elementwise_max``,
``__builtin_elementwise_min``, ``__builtin_elementwise_abs``,
-``__builtin_elementwise_ctlz``, ``__builtin_elementwise_cttz``, and
+``__builtin_elementwise_clzg``, ``__builtin_elementwise_ctzg``, and
``__builtin_elementwise_fma`` can be called in a ``constexpr`` context.
No implicit promotion of integer types takes place. The mixing of integer types
@@ -884,11 +884,11 @@ T __builtin_elementwise_fshr(T x, T y, T z) perform a
funnel shift right. Co
right by z (modulo the bit
width of the original arguments),
and the least significant bits
are extracted to produce
a result that is the same size
as the original arguments.
- T __builtin_elementwise_ctlz(T x[, T y]) return the number of leading 0
bits in the first argument. If integer types
+ T __builtin_elementwise_clzg(T x[, T y]) return the number of leading 0
bits in the first argument. If integer types
the first argument is 0 and an
optional second argument is provided,
the second argument is
returned. It is undefined behaviour if the
first argument is 0 and no
second argument is provided.
- T __builtin_elementwise_cttz(T x[, T y]) return the number of trailing
0 bits in the first argument. If integer types
+ T __builtin_elementwise_ctzg(T x[, T y]) return the number of trailing
0 bits in the first argument. If integer types
the first argument is 0 and an
optional second argument is provided,
the second argument is
returned. It is undefined behaviour if the
first argument is 0 and no
second argument is provided.
diff --git a/clang/include/clang/Basic/Builtins.td
b/clang/include/clang/Basic/Builtins.td
index 27fc6f008d743..1111cfacb8559 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -1551,13 +1551,13 @@ def ElementwiseFshr : Builtin {
}
def ElementwiseCtlz : Builtin {
- let Spellings = ["__builtin_elementwise_ctlz"];
+ let Spellings = ["__builtin_elementwise_clzg"];
let Attributes = [NoThrow, Const, CustomTypeChecking, Constexpr];
let Prototype = "void(...)";
}
def ElementwiseCttz : Builtin {
- let Spellings = ["__builtin_elementwise_cttz"];
+ let Spellings = ["__builtin_elementwise_ctzg"];
let Attributes = [NoThrow, Const, CustomTypeChecking, Constexpr];
let Prototype = "void(...)";
}
diff --git a/clang/include/clang/Basic/DiagnosticASTKinds.td
b/clang/include/clang/Basic/DiagnosticASTKinds.td
index a63bd80b89657..0be9146f70364 100644
--- a/clang/include/clang/Basic/DiagnosticASTKinds.td
+++ b/clang/include/clang/Basic/DiagnosticASTKinds.td
@@ -401,7 +401,7 @@ def note_constexpr_non_const_vectorelements : Note<
def note_constexpr_assumption_failed : Note<
"assumption evaluated to false">;
def note_constexpr_countzeroes_zero : Note<
- "evaluation of
%select{__builtin_elementwise_ctlz|__builtin_elementwise_cttz}0 "
+ "evaluation of
%select{__builtin_elementwise_clzg|__builtin_elementwise_ctzg}0 "
"with a zero value is undefined">;
def err_experimental_clang_interp_failed : Error<
"the experimental clang interpreter failed to evaluate an expression">;
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index ff6ef5a1f6864..57533db93d816 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -1831,7 +1831,7 @@ static bool
interp__builtin_elementwise_countzeroes(InterpState &S,
const CallExpr *Call,
unsigned BuiltinID) {
const bool HasZeroArg = Call->getNumArgs() == 2;
- const bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_cttz;
+ const bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_ctzg;
assert(Call->getNumArgs() == 1 || HasZeroArg);
if (Call->getArg(0)->getType()->isIntegerType()) {
PrimType ArgT = *S.getContext().classify(Call->getArg(0)->getType());
@@ -1855,7 +1855,7 @@ static bool
interp__builtin_elementwise_countzeroes(InterpState &S,
return false;
}
- if (BuiltinID == Builtin::BI__builtin_elementwise_ctlz) {
+ if (BuiltinID == Builtin::BI__builtin_elementwise_clzg) {
pushInteger(S, Val.countLeadingZeros(), Call->getType());
} else {
pushInteger(S, Val.countTrailingZeros(), Call->getType());
@@ -3164,8 +3164,8 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const
CallExpr *Call,
case Builtin::BI__builtin_ctzg:
return interp__builtin_ctz(S, OpPC, Frame, Call, BuiltinID);
- case Builtin::BI__builtin_elementwise_ctlz:
- case Builtin::BI__builtin_elementwise_cttz:
+ case Builtin::BI__builtin_elementwise_clzg:
+ case Builtin::BI__builtin_elementwise_ctzg:
return interp__builtin_elementwise_countzeroes(S, OpPC, Frame, Call,
BuiltinID);
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index 6c6909e5b2370..0479aeecb1a5e 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -12039,8 +12039,8 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr
*E) {
return Success(APValue(ResultElements.data(), ResultElements.size()), E);
}
- case Builtin::BI__builtin_elementwise_ctlz:
- case Builtin::BI__builtin_elementwise_cttz: {
+ case Builtin::BI__builtin_elementwise_clzg:
+ case Builtin::BI__builtin_elementwise_ctzg: {
APValue SourceLHS;
std::optional<APValue> Fallback;
if (!EvaluateAsRValue(Info, E->getArg(0), SourceLHS))
@@ -12064,19 +12064,19 @@ bool VectorExprEvaluator::VisitCallExpr(const
CallExpr *E) {
if (!Fallback) {
Info.FFDiag(E, diag::note_constexpr_countzeroes_zero)
<< /*IsTrailing=*/(E->getBuiltinCallee() ==
- Builtin::BI__builtin_elementwise_cttz);
+ Builtin::BI__builtin_elementwise_ctzg);
return false;
}
ResultElements.push_back(Fallback->getVectorElt(EltNum));
continue;
}
switch (E->getBuiltinCallee()) {
- case Builtin::BI__builtin_elementwise_ctlz:
+ case Builtin::BI__builtin_elementwise_clzg:
ResultElements.push_back(APValue(
APSInt(APInt(Info.Ctx.getIntWidth(DestEltTy), LHS.countl_zero()),
DestEltTy->isUnsignedIntegerOrEnumerationType())));
break;
- case Builtin::BI__builtin_elementwise_cttz:
+ case Builtin::BI__builtin_elementwise_ctzg:
ResultElements.push_back(APValue(
APSInt(APInt(Info.Ctx.getIntWidth(DestEltTy), LHS.countr_zero()),
DestEltTy->isUnsignedIntegerOrEnumerationType())));
@@ -13694,7 +13694,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
case Builtin::BI__builtin_clzll:
case Builtin::BI__builtin_clzs:
case Builtin::BI__builtin_clzg:
- case Builtin::BI__builtin_elementwise_ctlz:
+ case Builtin::BI__builtin_elementwise_clzg:
case Builtin::BI__lzcnt16: // Microsoft variants of count leading-zeroes
case Builtin::BI__lzcnt:
case Builtin::BI__lzcnt64: {
@@ -13710,7 +13710,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
std::optional<APSInt> Fallback;
if ((BuiltinOp == Builtin::BI__builtin_clzg ||
- BuiltinOp == Builtin::BI__builtin_elementwise_ctlz) &&
+ BuiltinOp == Builtin::BI__builtin_elementwise_clzg) &&
E->getNumArgs() > 1) {
APSInt FallbackTemp;
if (!EvaluateInteger(E->getArg(1), FallbackTemp, Info))
@@ -13729,7 +13729,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
BuiltinOp != Builtin::BI__lzcnt &&
BuiltinOp != Builtin::BI__lzcnt64;
- if (BuiltinOp == Builtin::BI__builtin_elementwise_ctlz) {
+ if (BuiltinOp == Builtin::BI__builtin_elementwise_clzg) {
Info.FFDiag(E, diag::note_constexpr_countzeroes_zero)
<< /*IsTrailing=*/false;
}
@@ -13789,7 +13789,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
case Builtin::BI__builtin_ctzll:
case Builtin::BI__builtin_ctzs:
case Builtin::BI__builtin_ctzg:
- case Builtin::BI__builtin_elementwise_cttz: {
+ case Builtin::BI__builtin_elementwise_ctzg: {
APSInt Val;
if (E->getArg(0)->getType()->isExtVectorBoolType()) {
APValue Vec;
@@ -13802,7 +13802,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
std::optional<APSInt> Fallback;
if ((BuiltinOp == Builtin::BI__builtin_ctzg ||
- BuiltinOp == Builtin::BI__builtin_elementwise_cttz) &&
+ BuiltinOp == Builtin::BI__builtin_elementwise_ctzg) &&
E->getNumArgs() > 1) {
APSInt FallbackTemp;
if (!EvaluateInteger(E->getArg(1), FallbackTemp, Info))
@@ -13814,7 +13814,7 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
if (Fallback)
return Success(*Fallback, E);
- if (BuiltinOp == Builtin::BI__builtin_elementwise_cttz) {
+ if (BuiltinOp == Builtin::BI__builtin_elementwise_ctzg) {
Info.FFDiag(E, diag::note_constexpr_countzeroes_zero)
<< /*IsTrailing=*/true;
}
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 172a521e63c17..c9c98e16fab43 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -3343,10 +3343,10 @@ RValue CodeGenFunction::EmitBuiltinExpr(const
GlobalDecl GD, unsigned BuiltinID,
case Builtin::BI__builtin_ctzl:
case Builtin::BI__builtin_ctzll:
case Builtin::BI__builtin_ctzg:
- case Builtin::BI__builtin_elementwise_cttz: {
+ case Builtin::BI__builtin_elementwise_ctzg: {
bool HasFallback =
(BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_ctzg ||
- BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_cttz) &&
+ BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_ctzg) &&
E->getNumArgs() > 1;
Value *ArgValue =
@@ -3360,7 +3360,7 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl
GD, unsigned BuiltinID,
// The elementwise builtins always exhibit zero-is-undef behaviour
Value *ZeroUndef = Builder.getInt1(
HasFallback || getTarget().isCLZForZeroUndef() ||
- BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_cttz);
+ BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_ctzg);
Value *Result = Builder.CreateCall(F, {ArgValue, ZeroUndef});
if (Result->getType() != ResultType)
Result =
@@ -3380,10 +3380,10 @@ RValue CodeGenFunction::EmitBuiltinExpr(const
GlobalDecl GD, unsigned BuiltinID,
case Builtin::BI__builtin_clzl:
case Builtin::BI__builtin_clzll:
case Builtin::BI__builtin_clzg:
- case Builtin::BI__builtin_elementwise_ctlz: {
+ case Builtin::BI__builtin_elementwise_clzg: {
bool HasFallback =
(BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_clzg ||
- BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_ctlz) &&
+ BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_clzg) &&
E->getNumArgs() > 1;
Value *ArgValue =
@@ -3397,7 +3397,7 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl
GD, unsigned BuiltinID,
// The elementwise builtins always exhibit zero-is-undef behaviour
Value *ZeroUndef = Builder.getInt1(
HasFallback || getTarget().isCLZForZeroUndef() ||
- BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_ctlz);
+ BuiltinIDIfNoAsmLabel == Builtin::BI__builtin_elementwise_clzg);
Value *Result = Builder.CreateCall(F, {ArgValue, ZeroUndef});
if (Result->getType() != ResultType)
Result =
diff --git a/clang/lib/Headers/avx512cdintrin.h
b/clang/lib/Headers/avx512cdintrin.h
index 39e76711ca7b3..6d572daaff11e 100644
--- a/clang/lib/Headers/avx512cdintrin.h
+++ b/clang/lib/Headers/avx512cdintrin.h
@@ -71,7 +71,7 @@ _mm512_maskz_conflict_epi32 (__mmask16 __U, __m512i __A)
static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR
_mm512_lzcnt_epi32(__m512i __A) {
- return (__m512i)__builtin_elementwise_ctlz((__v16si)__A,
+ return (__m512i)__builtin_elementwise_clzg((__v16si)__A,
(__v16si)_mm512_set1_epi32(32));
}
@@ -91,7 +91,7 @@ _mm512_maskz_lzcnt_epi32(__mmask16 __U, __m512i __A) {
static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR
_mm512_lzcnt_epi64(__m512i __A) {
- return (__m512i)__builtin_elementwise_ctlz(
+ return (__m512i)__builtin_elementwise_clzg(
(__v8di)__A, (__v8di)_mm512_set1_epi64((long long)64));
}
diff --git a/clang/lib/Headers/avx512vlcdintrin.h
b/clang/lib/Headers/avx512vlcdintrin.h
index 8f42675ba9b5d..62656f794ce2c 100644
--- a/clang/lib/Headers/avx512vlcdintrin.h
+++ b/clang/lib/Headers/avx512vlcdintrin.h
@@ -146,7 +146,7 @@ _mm256_maskz_conflict_epi32 (__mmask8 __U, __m256i __A)
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR
_mm_lzcnt_epi32(__m128i __A) {
- return (__m128i)__builtin_elementwise_ctlz((__v4si)__A,
+ return (__m128i)__builtin_elementwise_clzg((__v4si)__A,
(__v4si)_mm_set1_epi32(32));
}
@@ -166,7 +166,7 @@ _mm_maskz_lzcnt_epi32(__mmask8 __U, __m128i __A) {
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR
_mm256_lzcnt_epi32(__m256i __A) {
- return (__m256i)__builtin_elementwise_ctlz((__v8si)__A,
+ return (__m256i)__builtin_elementwise_clzg((__v8si)__A,
(__v8si)_mm256_set1_epi32(32));
}
@@ -186,7 +186,7 @@ _mm256_maskz_lzcnt_epi32(__mmask8 __U, __m256i __A) {
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR
_mm_lzcnt_epi64(__m128i __A) {
- return (__m128i)__builtin_elementwise_ctlz(
+ return (__m128i)__builtin_elementwise_clzg(
(__v2di)__A, (__v2di)_mm_set1_epi64x((long long)64));
}
@@ -206,7 +206,7 @@ _mm_maskz_lzcnt_epi64(__mmask8 __U, __m128i __A) {
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR
_mm256_lzcnt_epi64(__m256i __A) {
- return (__m256i)__builtin_elementwise_ctlz(
+ return (__m256i)__builtin_elementwise_clzg(
(__v4di)__A, (__v4di)_mm256_set1_epi64x((long long)64));
}
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index 077f4311ed729..dc1d06894d48e 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -3181,8 +3181,8 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl,
unsigned BuiltinID,
TheCall->setType(Magnitude.get()->getType());
break;
}
- case Builtin::BI__builtin_elementwise_ctlz:
- case Builtin::BI__builtin_elementwise_cttz:
+ case Builtin::BI__builtin_elementwise_clzg:
+ case Builtin::BI__builtin_elementwise_ctzg:
// These builtins can be unary or binary. Note for empty calls we call the
// unary checker in order to not emit an error that says the function
// expects 2 arguments, which would be misleading.
diff --git a/clang/test/CodeGen/builtins-elementwise-math.c
b/clang/test/CodeGen/builtins-elementwise-math.c
index 188a6c3a30f0e..e9344d8fe0b8b 100644
--- a/clang/test/CodeGen/builtins-elementwise-math.c
+++ b/clang/test/CodeGen/builtins-elementwise-math.c
@@ -1266,98 +1266,98 @@ void test_builtin_elementwise_fshl(long long int i1,
long long int i2,
u4 tmp_vu_r = __builtin_elementwise_fshr(vu1, vu2, vu3);
}
-void test_builtin_elementwise_ctlz(si8 vs1, si8 vs2, u4 vu1,
+void test_builtin_elementwise_clzg(si8 vs1, si8 vs2, u4 vu1,
long long int lli, short si,
_BitInt(31) bi, int i,
char ci) {
// CHECK: [[V8S1:%.+]] = load <8 x i16>, ptr %vs1.addr
// CHECK-NEXT: call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> [[V8S1]], i1 true)
- vs1 = __builtin_elementwise_ctlz(vs1);
+ vs1 = __builtin_elementwise_clzg(vs1);
// CHECK: [[V8S1:%.+]] = load <8 x i16>, ptr %vs1.addr
// CHECK-NEXT: [[CLZ:%.+]] = call <8 x i16> @llvm.ctlz.v8i16(<8 x i16>
[[V8S1]], i1 true)
// CHECK-NEXT: [[ISZERO:%.+]] = icmp eq <8 x i16> [[V8S1]], zeroinitializer
// CHECK-NEXT: [[V8S2:%.+]] = load <8 x i16>, ptr %vs2.addr
// select <8 x i1> [[ISZERO]], <8 x i16> [[CLZ]], <8 x i16> [[V8S2]]
- vs1 = __builtin_elementwise_ctlz(vs1, vs2);
+ vs1 = __builtin_elementwise_clzg(vs1, vs2);
// CHECK: [[V4U1:%.+]] = load <4 x i32>, ptr %vu1.addr
// CHECK-NEXT: call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> [[V4U1]], i1 true)
- vu1 = __builtin_elementwise_ctlz(vu1);
+ vu1 = __builtin_elementwise_clzg(vu1);
// CHECK: [[LLI:%.+]] = load i64, ptr %lli.addr
// CHECK-NEXT: call i64 @llvm.ctlz.i64(i64 [[LLI]], i1 true)
- lli = __builtin_elementwise_ctlz(lli);
+ lli = __builtin_elementwise_clzg(lli);
// CHECK: [[SI:%.+]] = load i16, ptr %si.addr
// CHECK-NEXT: call i16 @llvm.ctlz.i16(i16 [[SI]], i1 true)
- si = __builtin_elementwise_ctlz(si);
+ si = __builtin_elementwise_clzg(si);
// CHECK: [[BI1:%.+]] = load i32, ptr %bi.addr
// CHECK-NEXT: [[BI2:%.+]] = trunc i32 [[BI1]] to i31
// CHECK-NEXT: call i31 @llvm.ctlz.i31(i31 [[BI2]], i1 true)
- bi = __builtin_elementwise_ctlz(bi);
+ bi = __builtin_elementwise_clzg(bi);
// CHECK: [[BI1:%.+]] = load i32, ptr %bi.addr
// CHECK-NEXT: [[BI2:%.+]] = trunc i32 [[BI1]] to i31
// CHECK-NEXT: [[CLZ:%.+]] = call i31 @llvm.ctlz.i31(i31 [[BI2]], i1 true)
// CHECK-NEXT: [[ISZERO:%.+]] = icmp eq i31 [[BI2]], 0
// CHECK-NEXT: select i1 [[ISZERO]], i31 1, i31 [[CLZ]]
- bi = __builtin_elementwise_ctlz(bi, (_BitInt(31))1);
+ bi = __builtin_elementwise_clzg(bi, (_BitInt(31))1);
// CHECK: [[I:%.+]] = load i32, ptr %i.addr
// CHECK-NEXT: call i32 @llvm.ctlz.i32(i32 [[I]], i1 true)
- i = __builtin_elementwise_ctlz(i);
+ i = __builtin_elementwise_clzg(i);
// CHECK: [[CI:%.+]] = load i8, ptr %ci.addr
// CHECK-NEXT: call i8 @llvm.ctlz.i8(i8 [[CI]], i1 true)
- ci = __builtin_elementwise_ctlz(ci);
+ ci = __builtin_elementwise_clzg(ci);
}
-void test_builtin_elementwise_cttz(si8 vs1, si8 vs2, u4 vu1,
+void test_builtin_elementwise_ctzg(si8 vs1, si8 vs2, u4 vu1,
long long int lli, short si,
_BitInt(31) bi, int i,
char ci) {
// CHECK: [[V8S1:%.+]] = load <8 x i16>, ptr %vs1.addr
// CHECK-NEXT: call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[V8S1]], i1 true)
- vs1 = __builtin_elementwise_cttz(vs1);
+ vs1 = __builtin_elementwise_ctzg(vs1);
// CHECK: [[V8S1:%.+]] = load <8 x i16>, ptr %vs1.addr
// CHECK-NEXT: [[ctz:%.+]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16>
[[V8S1]], i1 true)
// CHECK-NEXT: [[ISZERO:%.+]] = icmp eq <8 x i16> [[V8S1]], zeroinitializer
// CHECK-NEXT: [[V8S2:%.+]] = load <8 x i16>, ptr %vs2.addr
// select <8 x i1> [[ISZERO]], <8 x i16> [[ctz]], <8 x i16> [[V8S2]]
- vs1 = __builtin_elementwise_cttz(vs1, vs2);
+ vs1 = __builtin_elementwise_ctzg(vs1, vs2);
// CHECK: [[V4U1:%.+]] = load <4 x i32>, ptr %vu1.addr
// CHECK-NEXT: call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[V4U1]], i1 true)
- vu1 = __builtin_elementwise_cttz(vu1);
+ vu1 = __builtin_elementwise_ctzg(vu1);
// CHECK: [[LLI:%.+]] = load i64, ptr %lli.addr
// CHECK-NEXT: call i64 @llvm.cttz.i64(i64 [[LLI]], i1 true)
- lli = __builtin_elementwise_cttz(lli);
+ lli = __builtin_elementwise_ctzg(lli);
// CHECK: [[SI:%.+]] = load i16, ptr %si.addr
// CHECK-NEXT: call i16 @llvm.cttz.i16(i16 [[SI]], i1 true)
- si = __builtin_elementwise_cttz(si);
+ si = __builtin_elementwise_ctzg(si);
// CHECK: [[BI1:%.+]] = load i32, ptr %bi.addr
// CHECK-NEXT: [[BI2:%.+]] = trunc i32 [[BI1]] to i31
// CHECK-NEXT: call i31 @llvm.cttz.i31(i31 [[BI2]], i1 true)
- bi = __builtin_elementwise_cttz(bi);
+ bi = __builtin_elementwise_ctzg(bi);
// CHECK: [[BI1:%.+]] = load i32, ptr %bi.addr
// CHECK-NEXT: [[BI2:%.+]] = trunc i32 [[BI1]] to i31
// CHECK-NEXT: [[ctz:%.+]] = call i31 @llvm.cttz.i31(i31 [[BI2]], i1 true)
// CHECK-NEXT: [[ISZERO:%.+]] = icmp eq i31 [[BI2]], 0
// CHECK-NEXT: select i1 [[ISZERO]], i31 1, i31 [[ctz]]
- bi = __builtin_elementwise_cttz(bi, (_BitInt(31))1);
+ bi = __builtin_elementwise_ctzg(bi, (_BitInt(31))1);
// CHECK: [[I:%.+]] = load i32, ptr %i.addr
// CHECK-NEXT: call i32 @llvm.cttz.i32(i32 [[I]], i1 true)
- i = __builtin_elementwise_cttz(i);
+ i = __builtin_elementwise_ctzg(i);
// CHECK: [[CI:%.+]] = load i8, ptr %ci.addr
// CHECK-NEXT: call i8 @llvm.cttz.i8(i8 [[CI]], i1 true)
- ci = __builtin_elementwise_cttz(ci);
+ ci = __builtin_elementwise_ctzg(ci);
}
diff --git a/clang/test/Sema/builtins-elementwise-math.c
b/clang/test/Sema/builtins-elementwise-math.c
index d6ce5c0a34746..f9df4a6f93e05 100644
--- a/clang/test/Sema/builtins-elementwise-math.c
+++ b/clang/test/Sema/builtins-elementwise-math.c
@@ -1357,46 +1357,46 @@ cfloat4 quux(cfloat4 x, float4 y) {
return __builtin_elementwise_pow(x,y);
}
-void test_builtin_elementwise_ctlz(int i32, int2 v2i32, short i16,
+void test_builtin_elementwise_clzg(int i32, int2 v2i32, short i16,
double f64, double2 v2f64) {
- f64 = __builtin_elementwise_ctlz(f64);
+ f64 = __builtin_elementwise_clzg(f64);
// expected-error@-1 {{1st argument must be a scalar or vector of integer
types (was 'double')}}
_Complex float c1;
- c1 = __builtin_elementwise_ctlz(c1);
+ c1 = __builtin_elementwise_clzg(c1);
// expected-error@-1 {{1st argument must be a scalar or vector of integer
types (was '_Complex float')}}
- v2i32 = __builtin_elementwise_ctlz(v2i32, i32);
+ v2i32 = __builtin_elementwise_clzg(v2i32, i32);
// expected-error@-1 {{arguments are of different types ('int2' (vector of 2
'int' values) vs 'int')}}
- v2i32 = __builtin_elementwise_ctlz(v2i32, f64);
+ v2i32 = __builtin_elementwise_clzg(v2i32, f64);
// expected-error@-1 {{arguments are of different types ('int2' (vector of 2
'int' values) vs 'double')}}
- v2i32 = __builtin_elementwise_ctlz();
+ v2i32 = __builtin_elementwise_clzg();
// expected-error@-1 {{too few arguments to function call, expected 1, have
0}}
- v2i32 = __builtin_elementwise_ctlz(v2i32, v2i32, f64);
+ v2i32 = __builtin_elementwise_clzg(v2i32, v2i32, f64);
// expected-error@-1 {{too many arguments to function call, expected 2, have
3}}
}
-void test_builtin_elementwise_cttz(int i32, int2 v2i32, short i16,
+void test_builtin_elementwise_ctzg(int i32, int2 v2i32, short i16,
double f64, double2 v2f64) {
- f64 = __builtin_elementwise_cttz(f64);
+ f64 = __builtin_elementwise_ctzg(f64);
// expected-error@-1 {{1st argument must be a scalar or vector of integer
types (was 'double')}}
_Complex float c1;
- c1 = __builtin_elementwise_cttz(c1);
+ c1 = __builtin_elementwise_ctzg(c1);
// expected-error@-1 {{1st argument must be a scalar or vector of integer
types (was '_Complex float')}}
- v2i32 = __builtin_elementwise_cttz(v2i32, i32);
+ v2i32 = __builtin_elementwise_ctzg(v2i32, i32);
// expected-error@-1 {{arguments are of different types ('int2' (vector of 2
'int' values) vs 'int')}}
- v2i32 = __builtin_elementwise_cttz(v2i32, f64);
+ v2i32 = __builtin_elementwise_ctzg(v2i32, f64);
// expected-error@-1 {{arguments are of different types ('int2' (vector of 2
'int' values) vs 'double')}}
- v2i32 = __builtin_elementwise_cttz();
+ v2i32 = __builtin_elementwise_ctzg();
// expected-error@-1 {{too few arguments to function call, expected 1, have
0}}
- v2i32 = __builtin_elementwise_cttz(v2i32, v2i32, f64);
+ v2i32 = __builtin_elementwise_ctzg(v2i32, v2i32, f64);
// expected-error@-1 {{too many arguments to function call, expected 2, have
3}}
}
diff --git a/clang/test/Sema/constant-builtins-vector.cpp
b/clang/test/Sema/constant-builtins-vector.cpp
index 714a7fb753214..7d6cd6d7e7ce7 100644
--- a/clang/test/Sema/constant-builtins-vector.cpp
+++ b/clang/test/Sema/constant-builtins-vector.cpp
@@ -895,47 +895,47 @@
CHECK_FOUR_FLOAT_VEC(__builtin_elementwise_abs((vector4double){-1.123, 2.123, -3
static_assert(__builtin_elementwise_abs((float)-1.123) - (float)1.123 < 1e-6);
// making sure one element works
#undef CHECK_FOUR_FLOAT_VEC
-static_assert(__builtin_elementwise_ctlz(2) == 30);
-static_assert(__builtin_elementwise_ctlz(2, 8) == 30);
-static_assert(__builtin_elementwise_ctlz(0, 8) == 8);
-static_assert(__builtin_elementwise_ctlz(0, 0) == 0);
-static_assert(__builtin_elementwise_ctlz((char)2) == 6);
-static_assert(__builtin_elementwise_ctlz((short)2) == 14);
-static_assert(__builtin_elementwise_ctlz((char)1) == 0x7);
-static_assert(__builtin_elementwise_ctlz((char)4) == 0x5);
-static_assert(__builtin_elementwise_ctlz((char)127) == 0x1);
-static_assert(__builtin_elementwise_ctlz((char)128) == 0x0);
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctlz((vector4char){1, 4, 127, (char)128})) == (LITTLE_END
? 0x00010507 : 0x07050100));
-
-constexpr int clz0 = __builtin_elementwise_ctlz(0);
+static_assert(__builtin_elementwise_clzg(2) == 30);
+static_assert(__builtin_elementwise_clzg(2, 8) == 30);
+static_assert(__builtin_elementwise_clzg(0, 8) == 8);
+static_assert(__builtin_elementwise_clzg(0, 0) == 0);
+static_assert(__builtin_elementwise_clzg((char)2) == 6);
+static_assert(__builtin_elementwise_clzg((short)2) == 14);
+static_assert(__builtin_elementwise_clzg((char)1) == 0x7);
+static_assert(__builtin_elementwise_clzg((char)4) == 0x5);
+static_assert(__builtin_elementwise_clzg((char)127) == 0x1);
+static_assert(__builtin_elementwise_clzg((char)128) == 0x0);
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_clzg((vector4char){1, 4, 127, (char)128})) == (LITTLE_END
? 0x00010507 : 0x07050100));
+
+constexpr int clz0 = __builtin_elementwise_clzg(0);
// expected-error@-1 {{must be initialized by a constant expression}} \
-// expected-note@-1 {{evaluation of __builtin_elementwise_ctlz with a zero
value is undefined}}
-constexpr vector4char clz1 = __builtin_elementwise_ctlz((vector4char){1, 0, 3,
4});
+// expected-note@-1 {{evaluation of __builtin_elementwise_clzg with a zero
value is undefined}}
+constexpr vector4char clz1 = __builtin_elementwise_clzg((vector4char){1, 0, 3,
4});
// expected-error@-1 {{must be initialized by a constant expression}} \
-// expected-note@-1 {{evaluation of __builtin_elementwise_ctlz with a zero
value is undefined}}
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctlz((vector4char){1, 0, 127, 0}, (vector4char){9, -1, 9,
-2})) == (LITTLE_END ? 0xFE01FF07 : 0x07FF01FE));
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctlz((vector4char){0, 0, 0, 0}, (vector4char){0, 0, 0,
0})) == 0);
-
-static_assert(__builtin_elementwise_cttz(2) == 1);
-static_assert(__builtin_elementwise_cttz(2, 8) == 1);
-static_assert(__builtin_elementwise_cttz(0, 8) == 8);
-static_assert(__builtin_elementwise_cttz(0, 0) == 0);
-static_assert(__builtin_elementwise_cttz((char)2) == 1);
-static_assert(__builtin_elementwise_cttz((short)2) == 1);
-static_assert(__builtin_elementwise_cttz((char)8) == 0x3);
-static_assert(__builtin_elementwise_cttz((char)32) == 0x5);
-static_assert(__builtin_elementwise_cttz((char)127) == 0x0);
-static_assert(__builtin_elementwise_cttz((char)128) == 0x7);
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_cttz((vector4char){8, 32, 127, (char)128})) ==
(LITTLE_END ? 0x07000503 : 0x03050007));
-
-constexpr int ctz0 = __builtin_elementwise_cttz(0);
+// expected-note@-1 {{evaluation of __builtin_elementwise_clzg with a zero
value is undefined}}
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_clzg((vector4char){1, 0, 127, 0}, (vector4char){9, -1, 9,
-2})) == (LITTLE_END ? 0xFE01FF07 : 0x07FF01FE));
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_clzg((vector4char){0, 0, 0, 0}, (vector4char){0, 0, 0,
0})) == 0);
+
+static_assert(__builtin_elementwise_ctzg(2) == 1);
+static_assert(__builtin_elementwise_ctzg(2, 8) == 1);
+static_assert(__builtin_elementwise_ctzg(0, 8) == 8);
+static_assert(__builtin_elementwise_ctzg(0, 0) == 0);
+static_assert(__builtin_elementwise_ctzg((char)2) == 1);
+static_assert(__builtin_elementwise_ctzg((short)2) == 1);
+static_assert(__builtin_elementwise_ctzg((char)8) == 0x3);
+static_assert(__builtin_elementwise_ctzg((char)32) == 0x5);
+static_assert(__builtin_elementwise_ctzg((char)127) == 0x0);
+static_assert(__builtin_elementwise_ctzg((char)128) == 0x7);
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctzg((vector4char){8, 32, 127, (char)128})) ==
(LITTLE_END ? 0x07000503 : 0x03050007));
+
+constexpr int ctz0 = __builtin_elementwise_ctzg(0);
// expected-error@-1 {{must be initialized by a constant expression}} \
-// expected-note@-1 {{evaluation of __builtin_elementwise_cttz with a zero
value is undefined}}
-constexpr vector4char ctz1 = __builtin_elementwise_cttz((vector4char){1, 0, 3,
4});
+// expected-note@-1 {{evaluation of __builtin_elementwise_ctzg with a zero
value is undefined}}
+constexpr vector4char ctz1 = __builtin_elementwise_ctzg((vector4char){1, 0, 3,
4});
// expected-error@-1 {{must be initialized by a constant expression}} \
-// expected-note@-1 {{evaluation of __builtin_elementwise_cttz with a zero
value is undefined}}
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_cttz((vector4char){8, 0, 127, 0}, (vector4char){9, -1, 9,
-2})) == (LITTLE_END ? 0xFE00FF03 : 0x03FF00FE));
-static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_cttz((vector4char){0, 0, 0, 0}, (vector4char){0, 0, 0,
0})) == 0);
+// expected-note@-1 {{evaluation of __builtin_elementwise_ctzg with a zero
value is undefined}}
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctzg((vector4char){8, 0, 127, 0}, (vector4char){9, -1, 9,
-2})) == (LITTLE_END ? 0xFE00FF03 : 0x03FF00FE));
+static_assert(__builtin_bit_cast(unsigned,
__builtin_elementwise_ctzg((vector4char){0, 0, 0, 0}, (vector4char){0, 0, 0,
0})) == 0);
// Non-vector floating point types.
static_assert(__builtin_elementwise_fma(2.0, 3.0, 4.0) == 10.0);
>From a04347c8c7bbd415d537654a8cda7644fd59227b Mon Sep 17 00:00:00 2001
From: Joseph Huber <hube...@outlook.com>
Date: Sun, 7 Sep 2025 16:38:28 -0500
Subject: [PATCH 2/2] OpenCL
---
libclc/clc/lib/generic/integer/clc_clz.inc | 2 +-
libclc/clc/lib/generic/integer/clc_ctz.inc | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/libclc/clc/lib/generic/integer/clc_clz.inc
b/libclc/clc/lib/generic/integer/clc_clz.inc
index ba7eb77d50036..1972a338fa16a 100644
--- a/libclc/clc/lib/generic/integer/clc_clz.inc
+++ b/libclc/clc/lib/generic/integer/clc_clz.inc
@@ -7,5 +7,5 @@
//===----------------------------------------------------------------------===//
_CLC_OVERLOAD _CLC_DEF __CLC_GENTYPE __clc_clz(__CLC_GENTYPE x) {
- return __builtin_elementwise_ctlz(x, (__CLC_GENTYPE)__CLC_GENSIZE);
+ return __builtin_elementwise_clzg(x, (__CLC_GENTYPE)__CLC_GENSIZE);
}
diff --git a/libclc/clc/lib/generic/integer/clc_ctz.inc
b/libclc/clc/lib/generic/integer/clc_ctz.inc
index cebd392542be9..b868482fdb5ae 100644
--- a/libclc/clc/lib/generic/integer/clc_ctz.inc
+++ b/libclc/clc/lib/generic/integer/clc_ctz.inc
@@ -7,5 +7,5 @@
//===----------------------------------------------------------------------===//
_CLC_OVERLOAD _CLC_DEF __CLC_GENTYPE __clc_ctz(__CLC_GENTYPE x) {
- return __builtin_elementwise_cttz(x, (__CLC_GENTYPE)__CLC_GENSIZE);
+ return __builtin_elementwise_ctzg(x, (__CLC_GENTYPE)__CLC_GENSIZE);
}
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
