https://github.com/Priyanshu3820 updated https://github.com/llvm/llvm-project/pull/171615
>From ea26e94f1ab429798860705ddd65d4e8f58c4103 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Wed, 10 Dec 2025 18:34:52 +0530 Subject: [PATCH 01/10] Implement convert_half builtin --- clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp | 41 +++++++-- .../CodeGenBuiltins/X86/cir-convert-half.c | 84 +++++++++++++++++++ 2 files changed, 119 insertions(+), 6 deletions(-) create mode 100644 clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c diff --git a/clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp b/clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp index f9e1f75a51143..852a7b9299e96 100644 --- a/clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp @@ -1662,12 +1662,41 @@ CIRGenFunction::emitX86BuiltinExpr(unsigned builtinID, const CallExpr *expr) { case X86::BI__builtin_ia32_cmpnltsd: case X86::BI__builtin_ia32_cmpnlesd: case X86::BI__builtin_ia32_cmpordsd: - case X86::BI__builtin_ia32_vcvtph2ps_mask: - case X86::BI__builtin_ia32_vcvtph2ps256_mask: - case X86::BI__builtin_ia32_vcvtph2ps512_mask: - case X86::BI__builtin_ia32_cvtneps2bf16_128_mask: - case X86::BI__builtin_ia32_cvtneps2bf16_256_mask: - case X86::BI__builtin_ia32_cvtneps2bf16_512_mask: + cgm.errorNYI(expr->getSourceRange(), + std::string("unimplemented X86 builtin call: ") + + getContext().BuiltinInfo.getName(builtinID)); + return {}; + case X86::BI__builtin_ia32_vcvtph2ps_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, "x86.avx512.mask.vcvtph2ps.128", + convertType(expr->getType()), ops); + } + case X86::BI__builtin_ia32_vcvtph2ps256_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, "x86.avx512.mask.vcvtph2ps.256", + convertType(expr->getType()), ops); + } + case X86::BI__builtin_ia32_vcvtph2ps512_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, "x86.avx512.mask.vcvtph2ps.512", + convertType(expr->getType()), ops); + } + case X86::BI__builtin_ia32_cvtneps2bf16_128_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, + "x86.avx512bf16.mask.cvtneps2bf16.128", + convertType(expr->getType()), ops); + } + case X86::BI__builtin_ia32_cvtneps2bf16_256_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, "x86.avx512bf16.cvtneps2bf16.256", + convertType(expr->getType()), ops); + } + case X86::BI__builtin_ia32_cvtneps2bf16_512_mask: { + mlir::Location loc = getLoc(expr->getExprLoc()); + return emitIntrinsicCallOp(builder, loc, "x86.avx512bf16.cvtneps2bf16.512", + convertType(expr->getType()), ops); + } case X86::BI__cpuid: case X86::BI__cpuidex: case X86::BI__emul: diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c new file mode 100644 index 0000000000000..8c6837a2e4e4a --- /dev/null +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -0,0 +1,84 @@ +// Test X86-specific convert_half builtins + +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +typedef float __m512 __attribute__((__vector_size__(64), __aligned__(64))); +typedef float __m256 __attribute__((__vector_size__(32), __aligned__(32))); +typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16))); +typedef int __m256i __attribute__((__vector_size__(32), __aligned__(32))); +typedef int __m128i __attribute__((__vector_size__(16), __aligned__(16))); +typedef int __mmask16; +typedef unsigned char __mmask8; +typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); +typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); + +// Test __builtin_ia32_vcvtph2ps512_mask +__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { + return __builtin_ia32_vcvtph2ps512_mask(a, src, k); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask +// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.512 +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps512_mask +// LLVM: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps512_mask +// OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 + +// Test __builtin_ia32_vcvtph2ps256_mask +__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps256_mask(a, src, k); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask +// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.256 +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps256_mask +// LLVM: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps256_mask +// OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 + +// Test __builtin_ia32_vcvtph2ps_mask +__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps_mask(a, src, k); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask +// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.128 +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps_mask +// LLVM: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps_mask +// OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 + +// Test __builtin_ia32_cvtneps2bf16_512_mask +__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { + return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask +// CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.512 +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask +// LLVM: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask +// OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 + +// Test __builtin_ia32_cvtneps2bf16_256_mask +__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask +// CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.256 +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask +// LLVM: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask +// OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 + +// Test __builtin_ia32_cvtneps2bf16_128_mask +__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask +// CIR: cir.call @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask +// LLVM: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask +// OGCG: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 \ No newline at end of file >From 342dc426223a3c70ebee91a35cff4e7f75972991 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 00:55:19 +0530 Subject: [PATCH 02/10] Update clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c --- .../CodeGenBuiltins/X86/cir-convert-half.c | 24 +++++++++---------- 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index 8c6837a2e4e4a..9f4571cc49d09 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -18,8 +18,8 @@ typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); // Test __builtin_ia32_vcvtph2ps512_mask -__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { - return __builtin_ia32_vcvtph2ps512_mask(a, src, k); +__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k, __m512 passthru) { + return __builtin_ia32_vcvtph2ps512_mask(a, src, k, passthru); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.512 @@ -29,8 +29,8 @@ __m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { // OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 // Test __builtin_ia32_vcvtph2ps256_mask -__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { - return __builtin_ia32_vcvtph2ps256_mask(a, src, k); +__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k, __m256 passthru) { + return __builtin_ia32_vcvtph2ps256_mask(a, src, k, passthru); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.256 @@ -40,8 +40,8 @@ __m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { // OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 // Test __builtin_ia32_vcvtph2ps_mask -__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { - return __builtin_ia32_vcvtph2ps_mask(a, src, k); +__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k, __m128 passthru) { + return __builtin_ia32_vcvtph2ps_mask(a, src, k, passthru); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.128 @@ -51,8 +51,8 @@ __m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { // OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 // Test __builtin_ia32_cvtneps2bf16_512_mask -__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { - return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); +__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u, __m256bh passthru) { + return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u, passthru); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask // CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.512 @@ -62,8 +62,8 @@ __m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { // OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 // Test __builtin_ia32_cvtneps2bf16_256_mask -__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { - return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); +__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u, __m128bh passthru) { + return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u, passthru); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask // CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.256 @@ -73,8 +73,8 @@ __m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { // OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 // Test __builtin_ia32_cvtneps2bf16_128_mask -__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { - return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); +__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u, __m128bh passthru) { + return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u, passthru); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask // CIR: cir.call @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 >From 0b40cb83ab0f4cb23872d95f69cfa1f4138ed123 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 01:15:13 +0530 Subject: [PATCH 03/10] Update test and fix formatting --- .../CIR/CodeGenBuiltins/X86/cir-convert-half.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index 9f4571cc49d09..e9c2404f48df6 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -40,8 +40,8 @@ __m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k, __m256 passthru // OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 // Test __builtin_ia32_vcvtph2ps_mask -__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k, __m128 passthru) { - return __builtin_ia32_vcvtph2ps_mask(a, src, k, passthru); +__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps_mask(a, src, k); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.128 @@ -51,8 +51,8 @@ __m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k, __m128 passthru) { // OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 // Test __builtin_ia32_cvtneps2bf16_512_mask -__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u, __m256bh passthru) { - return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u, passthru); +__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { + return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask // CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.512 @@ -62,8 +62,8 @@ __m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u, __m256bh // OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 // Test __builtin_ia32_cvtneps2bf16_256_mask -__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u, __m128bh passthru) { - return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u, passthru); +__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask // CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.256 @@ -73,8 +73,8 @@ __m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u, __m128bh p // OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 // Test __builtin_ia32_cvtneps2bf16_128_mask -__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u, __m128bh passthru) { - return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u, passthru); +__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); } // CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask // CIR: cir.call @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 >From da4f0523b38f0f596fbc53b8cdb31b7d265e62fa Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 01:38:55 +0530 Subject: [PATCH 04/10] Update clang\test\CIR\CodeGenBuiltins\X86\cir-convert-half.c --- clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index e9c2404f48df6..8c6837a2e4e4a 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -18,8 +18,8 @@ typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); // Test __builtin_ia32_vcvtph2ps512_mask -__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k, __m512 passthru) { - return __builtin_ia32_vcvtph2ps512_mask(a, src, k, passthru); +__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { + return __builtin_ia32_vcvtph2ps512_mask(a, src, k); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.512 @@ -29,8 +29,8 @@ __m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k, __m512 passthr // OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 // Test __builtin_ia32_vcvtph2ps256_mask -__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k, __m256 passthru) { - return __builtin_ia32_vcvtph2ps256_mask(a, src, k, passthru); +__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps256_mask(a, src, k); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.256 >From 4b3f3415835a7c682ce5b27ddae7f90963919a00 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 09:02:57 +0530 Subject: [PATCH 05/10] Update clang\test\CIR\CodeGenBuiltins\X86\cir-convert-half.c --- clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index 8c6837a2e4e4a..5e5060e1da1bf 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -18,8 +18,8 @@ typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); // Test __builtin_ia32_vcvtph2ps512_mask -__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { - return __builtin_ia32_vcvtph2ps512_mask(a, src, k); +__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k, _m512 dst) { + return __builtin_ia32_vcvtph2ps512_mask(a, src, k, dst); } // CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask // CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.512 >From 584276b2804ba38f3fd7be7fbc01e51e54c31f6d Mon Sep 17 00:00:00 2001 From: Priyanshu Kumar <[email protected]> Date: Thu, 11 Dec 2025 06:16:55 +0000 Subject: [PATCH 06/10] Update test --- .../CodeGenBuiltins/X86/cir-convert-half.c | 166 +++++++++--------- 1 file changed, 83 insertions(+), 83 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index 5e5060e1da1bf..09c08ef06a133 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -1,84 +1,84 @@ -// Test X86-specific convert_half builtins - -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-cir %s -o %t.cir -// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-llvm %s -o %t-cir.ll -// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o %t.ll -// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s - -typedef float __m512 __attribute__((__vector_size__(64), __aligned__(64))); -typedef float __m256 __attribute__((__vector_size__(32), __aligned__(32))); -typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16))); -typedef int __m256i __attribute__((__vector_size__(32), __aligned__(32))); -typedef int __m128i __attribute__((__vector_size__(16), __aligned__(16))); -typedef int __mmask16; -typedef unsigned char __mmask8; -typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); -typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); - -// Test __builtin_ia32_vcvtph2ps512_mask -__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k, _m512 dst) { - return __builtin_ia32_vcvtph2ps512_mask(a, src, k, dst); -} -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask -// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.512 -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps512_mask -// LLVM: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps512_mask -// OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 - -// Test __builtin_ia32_vcvtph2ps256_mask -__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { - return __builtin_ia32_vcvtph2ps256_mask(a, src, k); -} -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask -// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.256 -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps256_mask -// LLVM: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps256_mask -// OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 - -// Test __builtin_ia32_vcvtph2ps_mask -__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { - return __builtin_ia32_vcvtph2ps_mask(a, src, k); -} -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask -// CIR: cir.call @llvm.x86.avx512.mask.vcvtph2ps.128 -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps_mask -// LLVM: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps_mask -// OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 - -// Test __builtin_ia32_cvtneps2bf16_512_mask -__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { - return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); -} -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask -// CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.512 -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask -// LLVM: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask -// OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 - -// Test __builtin_ia32_cvtneps2bf16_256_mask -__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { - return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); -} -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask -// CIR: cir.call @llvm.x86.avx512bf16.cvtneps2bf16.256 -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask -// LLVM: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask -// OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 - -// Test __builtin_ia32_cvtneps2bf16_128_mask -__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { - return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); -} -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask -// CIR: cir.call @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask -// LLVM: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask +// Test X86-specific convert_half builtins + +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +typedef float __m512 __attribute__((__vector_size__(64), __aligned__(64))); +typedef float __m256 __attribute__((__vector_size__(32), __aligned__(32))); +typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16))); +typedef int __m256i __attribute__((__vector_size__(32), __aligned__(32))); +typedef int __m128i __attribute__((__vector_size__(16), __aligned__(16))); +typedef int __mmask16; +typedef unsigned char __mmask8; +typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); +typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); + +// Test __builtin_ia32_vcvtph2ps512_mask +__m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { + return __builtin_ia32_vcvtph2ps512_mask(a, src, k, 4); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.512" +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps512_mask +// LLVM: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps512_mask +// OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 + +// Test __builtin_ia32_vcvtph2ps256_mask +__m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps256_mask(a, src, k); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.256" +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps256_mask +// LLVM: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps256_mask +// OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 + +// Test __builtin_ia32_vcvtph2ps_mask +__m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { + return __builtin_ia32_vcvtph2ps_mask(a, src, k); +} +// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.128" +// LLVM-LABEL: define {{.*}} @test_vcvtph2ps_mask +// LLVM: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 +// OGCG-LABEL: define {{.*}} @test_vcvtph2ps_mask +// OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 + +// Test __builtin_ia32_cvtneps2bf16_512_mask +__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { + return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.512" +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask +// LLVM: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask +// OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 + +// Test __builtin_ia32_cvtneps2bf16_256_mask +__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.256" +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask +// LLVM: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask +// OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 + +// Test __builtin_ia32_cvtneps2bf16_128_mask +__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); +} +// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.mask.cvtneps2bf16.128" +// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask +// LLVM: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 +// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask // OGCG: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 \ No newline at end of file >From 16ba37f08b3d70d6899dc2c600d0bc268dded41e Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 12:55:15 +0530 Subject: [PATCH 07/10] Update test and lit.cfg.py --- clang/test/lit.cfg.py | 890 +++++++++++++++++++++--------------------- 1 file changed, 446 insertions(+), 444 deletions(-) diff --git a/clang/test/lit.cfg.py b/clang/test/lit.cfg.py index 52b275c095475..093400b0d708a 100644 --- a/clang/test/lit.cfg.py +++ b/clang/test/lit.cfg.py @@ -1,444 +1,446 @@ -# -*- Python -*- - -import os -import platform -import re -import subprocess -import tempfile - -import lit.formats -import lit.util - -from lit.llvm import llvm_config -from lit.llvm.subst import ToolSubst -from lit.llvm.subst import FindTool - -# Configuration file for the 'lit' test runner. - -# name: The name of this test suite. -config.name = "Clang" - -# TODO: Consolidate the logic for turning on the internal shell by default for all LLVM test suites. -# See https://github.com/llvm/llvm-project/issues/106636 for more details. -# -# We prefer the lit internal shell which provides a better user experience on failures -# and is faster unless the user explicitly disables it with LIT_USE_INTERNAL_SHELL=0 -# env var. -use_lit_shell = True -lit_shell_env = os.environ.get("LIT_USE_INTERNAL_SHELL") -if lit_shell_env: - use_lit_shell = lit.util.pythonize_bool(lit_shell_env) - -# testFormat: The test format to use to interpret tests. -# -# For now we require '&&' between commands, until they get globally killed and -# the test runner updated. -config.test_format = lit.formats.ShTest(execute_external=not use_lit_shell) - -# suffixes: A list of file extensions to treat as test files. -config.suffixes = [ - ".c", - ".cpp", - ".i", - ".cir", - ".cppm", - ".m", - ".mm", - ".cu", - ".cuh", - ".hip", - ".hlsl", - ".ll", - ".cl", - ".clcpp", - ".s", - ".S", - ".modulemap", - ".test", - ".rs", - ".ifs", - ".rc", -] - -# excludes: A list of directories to exclude from the testsuite. The 'Inputs' -# subdirectories contain auxiliary inputs for various tests in their parent -# directories. -config.excludes = [ - "Inputs", - "CMakeLists.txt", - "README.txt", - "LICENSE.txt", - "debuginfo-tests", -] - -# test_source_root: The root path where tests are located. -config.test_source_root = os.path.dirname(__file__) - -# test_exec_root: The root path where tests should be run. -config.test_exec_root = os.path.join(config.clang_obj_root, "test") - -llvm_config.use_default_substitutions() - -llvm_config.use_clang() - -config.substitutions.append(("%src_dir", config.clang_src_dir)) - -config.substitutions.append(("%src_include_dir", config.clang_src_dir + "/include")) - -config.substitutions.append(("%target_triple", config.target_triple)) - -config.substitutions.append(("%PATH%", config.environment["PATH"])) - - -# For each occurrence of a clang tool name, replace it with the full path to -# the build directory holding that tool. We explicitly specify the directories -# to search to ensure that we get the tools just built and not some random -# tools that might happen to be in the user's PATH. -tool_dirs = [config.clang_tools_dir, config.llvm_tools_dir] - -tools = [ - "apinotes-test", - "c-index-test", - "cir-opt", - "clang-diff", - "clang-format", - "clang-repl", - "llvm-offload-binary", - "clang-tblgen", - "clang-scan-deps", - "clang-installapi", - "opt", - "llvm-ifs", - "yaml2obj", - "clang-linker-wrapper", - "clang-nvlink-wrapper", - "clang-sycl-linker", - "llvm-lto", - "llvm-lto2", - "llvm-profdata", - "llvm-readtapi", - ToolSubst( - "%clang_extdef_map", - command=FindTool("clang-extdef-mapping"), - unresolved="ignore", - ), -] - -if config.clang_examples: - config.available_features.add("examples") - - -def have_host_out_of_process_jit_feature_support(): - clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) - - if not clang_repl_exe: - return False - - testcode = b"\n".join([b"int i = 0;", b"%quit"]) - - try: - clang_repl_cmd = subprocess.run( - [clang_repl_exe, "-orc-runtime", "-oop-executor"], - stdout=subprocess.PIPE, - stderr=subprocess.PIPE, - input=testcode, - ) - except OSError: - return False - - if clang_repl_cmd.returncode == 0: - return True - - return False - - -def run_clang_repl(args): - clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) - - if not clang_repl_exe: - return "" - - try: - clang_repl_cmd = subprocess.Popen( - [clang_repl_exe, args], stdout=subprocess.PIPE - ) - except OSError: - print("could not exec clang-repl") - return "" - - clang_repl_out = clang_repl_cmd.stdout.read().decode("ascii") - clang_repl_cmd.wait() - - return clang_repl_out - - -def have_host_jit_feature_support(feature_name): - return "true" in run_clang_repl("--host-supports-" + feature_name) - -def have_host_clang_repl_cuda(): - clang_repl_exe = lit.util.which('clang-repl', config.clang_tools_dir) - - if not clang_repl_exe: - return False - - testcode = b'\n'.join([ - b"__global__ void test_func() {}", - b"test_func<<<1,1>>>();", - b"extern \"C\" int puts(const char *s);", - b"puts(cudaGetLastError() ? \"failure\" : \"success\");", - b"%quit" - ]) - try: - clang_repl_cmd = subprocess.run([clang_repl_exe, '--cuda'], - stdout=subprocess.PIPE, - stderr=subprocess.PIPE, - input=testcode) - except OSError: - return False - - if clang_repl_cmd.returncode == 0: - if clang_repl_cmd.stdout.find(b"success") != -1: - return True - - return False - -if have_host_jit_feature_support('jit'): - config.available_features.add('host-supports-jit') - - if have_host_clang_repl_cuda(): - config.available_features.add('host-supports-cuda') - hosttriple = run_clang_repl("--host-jit-triple") - config.substitutions.append(("%host-jit-triple", hosttriple.strip())) - - if have_host_out_of_process_jit_feature_support(): - config.available_features.add("host-supports-out-of-process-jit") - -if config.clang_staticanalyzer: - config.available_features.add("staticanalyzer") - tools.append("clang-check") - - if config.clang_staticanalyzer_z3: - config.available_features.add("z3") - if config.clang_staticanalyzer_z3_mock: - config.available_features.add("z3-mock") - else: - config.available_features.add("no-z3") - - check_analyzer_fixit_path = os.path.join( - config.test_source_root, "Analysis", "check-analyzer-fixit.py" - ) - config.substitutions.append( - ( - "%check_analyzer_fixit", - '"%s" %s' % (config.python_executable, check_analyzer_fixit_path), - ) - ) - - csv2json_path = os.path.join(config.test_source_root, "Analysis", "csv2json.py") - config.substitutions.append( - ( - "%csv2json", - '"%s" %s' % (config.python_executable, csv2json_path), - ) - ) - -# ClangIR support -if config.clang_enable_cir: - config.available_features.add("cir-support") - -llvm_config.add_tool_substitutions(tools, tool_dirs) - -config.substitutions.append( - ( - "%hmaptool", - "'%s' %s" - % ( - config.python_executable, - os.path.join(config.clang_src_dir, "utils", "hmaptool", "hmaptool"), - ), - ) -) - -config.substitutions.append( - ( - "%deps-to-rsp", - '"%s" %s' - % ( - config.python_executable, - os.path.join(config.clang_src_dir, "utils", "module-deps-to-rsp.py"), - ), - ) -) - -# Determine whether the test target is compatible with execution on the host. -if "aarch64" in config.host_arch: - config.available_features.add("aarch64-host") - -# Some tests are sensitive to whether clang is statically or dynamically linked -# to other libraries. -if not (config.build_shared_libs or config.link_llvm_dylib or config.link_clang_dylib): - config.available_features.add("static-libs") - -# Plugins (loadable modules) -if config.has_plugins and config.llvm_plugin_ext: - config.available_features.add("plugins") - -if config.clang_default_pie_on_linux: - config.available_features.add("default-pie-on-linux") - -# Set available features we allow tests to conditionalize on. -# -if config.clang_default_cxx_stdlib != "": - config.available_features.add( - "default-cxx-stdlib={}".format(config.clang_default_cxx_stdlib) - ) - -# As of 2011.08, crash-recovery tests still do not pass on FreeBSD. -if platform.system() not in ["FreeBSD"]: - config.available_features.add("crash-recovery") - -# ANSI escape sequences in non-dumb terminal -if platform.system() not in ["Windows"]: - config.available_features.add("ansi-escape-sequences") - -# Capability to print utf8 to the terminal. -# Windows expects codepage, unless Wide API. -if platform.system() not in ["Windows"]: - config.available_features.add("utf8-capable-terminal") - -# Support for libgcc runtime. Used to rule out tests that require -# clang to run with -rtlib=libgcc. -if platform.system() not in ["Darwin", "Fuchsia"]: - config.available_features.add("libgcc") - -# Case-insensitive file system - - -def is_filesystem_case_insensitive(): - os.makedirs(config.test_exec_root, exist_ok=True) - handle, path = tempfile.mkstemp(prefix="case-test", dir=config.test_exec_root) - isInsensitive = os.path.exists( - os.path.join(os.path.dirname(path), os.path.basename(path).upper()) - ) - os.close(handle) - os.remove(path) - return isInsensitive - - -if is_filesystem_case_insensitive(): - config.available_features.add("case-insensitive-filesystem") - -# Tests that require the /dev/fd filesystem. -if os.path.exists("/dev/fd/0") and sys.platform not in ["cygwin"]: - config.available_features.add("dev-fd-fs") - -# Set on native MS environment. -if re.match(r".*-(windows-msvc)$", config.target_triple): - config.available_features.add("ms-sdk") - -# [PR8833] LLP64-incompatible tests -if not re.match( - r"^(aarch64|arm64ec|x86_64).*-(windows-msvc|windows-gnu)$", config.target_triple -): - config.available_features.add("LP64") - -# Tests that are specific to the Apple Silicon macOS. -if re.match(r"^arm64(e)?-apple-(macos|darwin)", config.target_triple): - config.available_features.add("apple-silicon-mac") - -# [PR18856] Depends to remove opened file. On win32, a file could be removed -# only if all handles were closed. -if platform.system() not in ["Windows"]: - config.available_features.add("can-remove-opened-file") - -# Features -known_arches = ["x86_64", "mips64", "ppc64", "aarch64"] -if any(config.target_triple.startswith(x) for x in known_arches): - config.available_features.add("clang-target-64-bits") - - -def calculate_arch_features(arch_string): - features = [] - for arch in arch_string.split(): - features.append(arch.lower() + "-registered-target") - return features - - -llvm_config.feature_config( - [ - ("--assertion-mode", {"ON": "asserts"}), - ("--cxxflags", {r"-D_GLIBCXX_DEBUG\b": "libstdcxx-safe-mode"}), - ("--targets-built", calculate_arch_features), - ] -) - -if lit.util.which("xmllint"): - config.available_features.add("xmllint") - -if config.enable_backtrace: - config.available_features.add("backtrace") - -if config.enable_threads: - config.available_features.add("thread_support") - -# Check if we should allow outputs to console. -run_console_tests = int(lit_config.params.get("enable_console", "0")) -if run_console_tests != 0: - config.available_features.add("console") - -lit.util.usePlatformSdkOnDarwin(config, lit_config) -macOSSDKVersion = lit.util.findPlatformSdkVersionOnMacOS(config, lit_config) -if macOSSDKVersion is not None: - config.available_features.add("macos-sdk-" + str(macOSSDKVersion)) - -if os.path.exists("/etc/gentoo-release"): - config.available_features.add("gentoo") - -if config.enable_shared: - config.available_features.add("enable_shared") - -# Add a vendor-specific feature. -if config.clang_vendor_uti: - config.available_features.add("clang-vendor=" + config.clang_vendor_uti) - -if config.have_llvm_driver: - config.available_features.add("llvm-driver") - - -# Some tests perform deep recursion, which requires a larger pthread stack size -# than the relatively low default of 192 KiB for 64-bit processes on AIX. The -# `AIXTHREAD_STK` environment variable provides a non-intrusive way to request -# a larger pthread stack size for the tests. Various applications and runtime -# libraries on AIX use a default pthread stack size of 4 MiB, so we will use -# that as a default value here. -if "AIXTHREAD_STK" in os.environ: - config.environment["AIXTHREAD_STK"] = os.environ["AIXTHREAD_STK"] -elif platform.system() == "AIX": - config.environment["AIXTHREAD_STK"] = "4194304" - -# Some tools support an environment variable "OBJECT_MODE" on AIX OS, which -# controls the kind of objects they will support. If there is no "OBJECT_MODE" -# environment variable specified, the default behaviour is to support 32-bit -# objects only. In order to not affect most test cases, which expect to support -# 32-bit and 64-bit objects by default, set the environment variable -# "OBJECT_MODE" to "any" by default on AIX OS. - -if "system-aix" in config.available_features: - config.substitutions.append(("llvm-nm", "env OBJECT_MODE=any llvm-nm")) - config.substitutions.append(("llvm-ar", "env OBJECT_MODE=any llvm-ar")) - config.substitutions.append(("llvm-ranlib", "env OBJECT_MODE=any llvm-ranlib")) - -# It is not realistically possible to account for all options that could -# possibly be present in system and user configuration files, so disable -# default configs for the test runs. -config.environment["CLANG_NO_DEFAULT_CONFIG"] = "1" - -if lit_config.update_tests: - import sys - import os - - utilspath = os.path.join(config.llvm_src_root, "utils") - sys.path.append(utilspath) - from update_any_test_checks import utc_lit_plugin - - lit_config.test_updaters.append(utc_lit_plugin) +# -*- Python -*- + +import os +import platform +import re +import subprocess +import tempfile + +import lit.formats +import lit.util + +from lit.llvm import llvm_config +from lit.llvm.subst import ToolSubst +from lit.llvm.subst import FindTool + +# Configuration file for the 'lit' test runner. + +# name: The name of this test suite. +config.name = "Clang" + +# TODO: Consolidate the logic for turning on the internal shell by default for all LLVM test suites. +# See https://github.com/llvm/llvm-project/issues/106636 for more details. +# +# We prefer the lit internal shell which provides a better user experience on failures +# and is faster unless the user explicitly disables it with LIT_USE_INTERNAL_SHELL=0 +# env var. +use_lit_shell = True +lit_shell_env = os.environ.get("LIT_USE_INTERNAL_SHELL") +if lit_shell_env: + use_lit_shell = lit.util.pythonize_bool(lit_shell_env) + +# testFormat: The test format to use to interpret tests. +# +# For now we require '&&' between commands, until they get globally killed and +# the test runner updated. +config.test_format = lit.formats.ShTest(execute_external=not use_lit_shell) + +# suffixes: A list of file extensions to treat as test files. +config.suffixes = [ + ".c", + ".cpp", + ".i", + ".cir", + ".cppm", + ".m", + ".mm", + ".cu", + ".cuh", + ".hip", + ".hlsl", + ".ll", + ".cl", + ".clcpp", + ".s", + ".S", + ".modulemap", + ".test", + ".rs", + ".ifs", + ".rc", +] + +# excludes: A list of directories to exclude from the testsuite. The 'Inputs' +# subdirectories contain auxiliary inputs for various tests in their parent +# directories. +config.excludes = [ + "Inputs", + "CMakeLists.txt", + "README.txt", + "LICENSE.txt", + "debuginfo-tests", +] + +# test_source_root: The root path where tests are located. +config.test_source_root = os.path.dirname(__file__) + +# test_exec_root: The root path where tests should be run. +config.test_exec_root = os.path.join(config.clang_obj_root, "test") + +llvm_config.use_default_substitutions() + +llvm_config.use_clang() + +config.substitutions.append(('%clang_cir', os.path.join(config.clang_tools_dir, 'clang'))) + +config.substitutions.append(("%src_dir", config.clang_src_dir)) + +config.substitutions.append(("%src_include_dir", config.clang_src_dir + "/include")) + +config.substitutions.append(("%target_triple", config.target_triple)) + +config.substitutions.append(("%PATH%", config.environment["PATH"])) + + +# For each occurrence of a clang tool name, replace it with the full path to +# the build directory holding that tool. We explicitly specify the directories +# to search to ensure that we get the tools just built and not some random +# tools that might happen to be in the user's PATH. +tool_dirs = [config.clang_tools_dir, config.llvm_tools_dir] + +tools = [ + "apinotes-test", + "c-index-test", + "cir-opt", + "clang-diff", + "clang-format", + "clang-repl", + "llvm-offload-binary", + "clang-tblgen", + "clang-scan-deps", + "clang-installapi", + "opt", + "llvm-ifs", + "yaml2obj", + "clang-linker-wrapper", + "clang-nvlink-wrapper", + "clang-sycl-linker", + "llvm-lto", + "llvm-lto2", + "llvm-profdata", + "llvm-readtapi", + ToolSubst( + "%clang_extdef_map", + command=FindTool("clang-extdef-mapping"), + unresolved="ignore", + ), +] + +if config.clang_examples: + config.available_features.add("examples") + + +def have_host_out_of_process_jit_feature_support(): + clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) + + if not clang_repl_exe: + return False + + testcode = b"\n".join([b"int i = 0;", b"%quit"]) + + try: + clang_repl_cmd = subprocess.run( + [clang_repl_exe, "-orc-runtime", "-oop-executor"], + stdout=subprocess.PIPE, + stderr=subprocess.PIPE, + input=testcode, + ) + except OSError: + return False + + if clang_repl_cmd.returncode == 0: + return True + + return False + + +def run_clang_repl(args): + clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) + + if not clang_repl_exe: + return "" + + try: + clang_repl_cmd = subprocess.Popen( + [clang_repl_exe, args], stdout=subprocess.PIPE + ) + except OSError: + print("could not exec clang-repl") + return "" + + clang_repl_out = clang_repl_cmd.stdout.read().decode("ascii") + clang_repl_cmd.wait() + + return clang_repl_out + + +def have_host_jit_feature_support(feature_name): + return "true" in run_clang_repl("--host-supports-" + feature_name) + +def have_host_clang_repl_cuda(): + clang_repl_exe = lit.util.which('clang-repl', config.clang_tools_dir) + + if not clang_repl_exe: + return False + + testcode = b'\n'.join([ + b"__global__ void test_func() {}", + b"test_func<<<1,1>>>();", + b"extern \"C\" int puts(const char *s);", + b"puts(cudaGetLastError() ? \"failure\" : \"success\");", + b"%quit" + ]) + try: + clang_repl_cmd = subprocess.run([clang_repl_exe, '--cuda'], + stdout=subprocess.PIPE, + stderr=subprocess.PIPE, + input=testcode) + except OSError: + return False + + if clang_repl_cmd.returncode == 0: + if clang_repl_cmd.stdout.find(b"success") != -1: + return True + + return False + +if have_host_jit_feature_support('jit'): + config.available_features.add('host-supports-jit') + + if have_host_clang_repl_cuda(): + config.available_features.add('host-supports-cuda') + hosttriple = run_clang_repl("--host-jit-triple") + config.substitutions.append(("%host-jit-triple", hosttriple.strip())) + + if have_host_out_of_process_jit_feature_support(): + config.available_features.add("host-supports-out-of-process-jit") + +if config.clang_staticanalyzer: + config.available_features.add("staticanalyzer") + tools.append("clang-check") + + if config.clang_staticanalyzer_z3: + config.available_features.add("z3") + if config.clang_staticanalyzer_z3_mock: + config.available_features.add("z3-mock") + else: + config.available_features.add("no-z3") + + check_analyzer_fixit_path = os.path.join( + config.test_source_root, "Analysis", "check-analyzer-fixit.py" + ) + config.substitutions.append( + ( + "%check_analyzer_fixit", + '"%s" %s' % (config.python_executable, check_analyzer_fixit_path), + ) + ) + + csv2json_path = os.path.join(config.test_source_root, "Analysis", "csv2json.py") + config.substitutions.append( + ( + "%csv2json", + '"%s" %s' % (config.python_executable, csv2json_path), + ) + ) + +# ClangIR support +if config.clang_enable_cir: + config.available_features.add("cir-support") + +llvm_config.add_tool_substitutions(tools, tool_dirs) + +config.substitutions.append( + ( + "%hmaptool", + "'%s' %s" + % ( + config.python_executable, + os.path.join(config.clang_src_dir, "utils", "hmaptool", "hmaptool"), + ), + ) +) + +config.substitutions.append( + ( + "%deps-to-rsp", + '"%s" %s' + % ( + config.python_executable, + os.path.join(config.clang_src_dir, "utils", "module-deps-to-rsp.py"), + ), + ) +) + +# Determine whether the test target is compatible with execution on the host. +if "aarch64" in config.host_arch: + config.available_features.add("aarch64-host") + +# Some tests are sensitive to whether clang is statically or dynamically linked +# to other libraries. +if not (config.build_shared_libs or config.link_llvm_dylib or config.link_clang_dylib): + config.available_features.add("static-libs") + +# Plugins (loadable modules) +if config.has_plugins and config.llvm_plugin_ext: + config.available_features.add("plugins") + +if config.clang_default_pie_on_linux: + config.available_features.add("default-pie-on-linux") + +# Set available features we allow tests to conditionalize on. +# +if config.clang_default_cxx_stdlib != "": + config.available_features.add( + "default-cxx-stdlib={}".format(config.clang_default_cxx_stdlib) + ) + +# As of 2011.08, crash-recovery tests still do not pass on FreeBSD. +if platform.system() not in ["FreeBSD"]: + config.available_features.add("crash-recovery") + +# ANSI escape sequences in non-dumb terminal +if platform.system() not in ["Windows"]: + config.available_features.add("ansi-escape-sequences") + +# Capability to print utf8 to the terminal. +# Windows expects codepage, unless Wide API. +if platform.system() not in ["Windows"]: + config.available_features.add("utf8-capable-terminal") + +# Support for libgcc runtime. Used to rule out tests that require +# clang to run with -rtlib=libgcc. +if platform.system() not in ["Darwin", "Fuchsia"]: + config.available_features.add("libgcc") + +# Case-insensitive file system + + +def is_filesystem_case_insensitive(): + os.makedirs(config.test_exec_root, exist_ok=True) + handle, path = tempfile.mkstemp(prefix="case-test", dir=config.test_exec_root) + isInsensitive = os.path.exists( + os.path.join(os.path.dirname(path), os.path.basename(path).upper()) + ) + os.close(handle) + os.remove(path) + return isInsensitive + + +if is_filesystem_case_insensitive(): + config.available_features.add("case-insensitive-filesystem") + +# Tests that require the /dev/fd filesystem. +if os.path.exists("/dev/fd/0") and sys.platform not in ["cygwin"]: + config.available_features.add("dev-fd-fs") + +# Set on native MS environment. +if re.match(r".*-(windows-msvc)$", config.target_triple): + config.available_features.add("ms-sdk") + +# [PR8833] LLP64-incompatible tests +if not re.match( + r"^(aarch64|arm64ec|x86_64).*-(windows-msvc|windows-gnu)$", config.target_triple +): + config.available_features.add("LP64") + +# Tests that are specific to the Apple Silicon macOS. +if re.match(r"^arm64(e)?-apple-(macos|darwin)", config.target_triple): + config.available_features.add("apple-silicon-mac") + +# [PR18856] Depends to remove opened file. On win32, a file could be removed +# only if all handles were closed. +if platform.system() not in ["Windows"]: + config.available_features.add("can-remove-opened-file") + +# Features +known_arches = ["x86_64", "mips64", "ppc64", "aarch64"] +if any(config.target_triple.startswith(x) for x in known_arches): + config.available_features.add("clang-target-64-bits") + + +def calculate_arch_features(arch_string): + features = [] + for arch in arch_string.split(): + features.append(arch.lower() + "-registered-target") + return features + + +llvm_config.feature_config( + [ + ("--assertion-mode", {"ON": "asserts"}), + ("--cxxflags", {r"-D_GLIBCXX_DEBUG\b": "libstdcxx-safe-mode"}), + ("--targets-built", calculate_arch_features), + ] +) + +if lit.util.which("xmllint"): + config.available_features.add("xmllint") + +if config.enable_backtrace: + config.available_features.add("backtrace") + +if config.enable_threads: + config.available_features.add("thread_support") + +# Check if we should allow outputs to console. +run_console_tests = int(lit_config.params.get("enable_console", "0")) +if run_console_tests != 0: + config.available_features.add("console") + +lit.util.usePlatformSdkOnDarwin(config, lit_config) +macOSSDKVersion = lit.util.findPlatformSdkVersionOnMacOS(config, lit_config) +if macOSSDKVersion is not None: + config.available_features.add("macos-sdk-" + str(macOSSDKVersion)) + +if os.path.exists("/etc/gentoo-release"): + config.available_features.add("gentoo") + +if config.enable_shared: + config.available_features.add("enable_shared") + +# Add a vendor-specific feature. +if config.clang_vendor_uti: + config.available_features.add("clang-vendor=" + config.clang_vendor_uti) + +if config.have_llvm_driver: + config.available_features.add("llvm-driver") + + +# Some tests perform deep recursion, which requires a larger pthread stack size +# than the relatively low default of 192 KiB for 64-bit processes on AIX. The +# `AIXTHREAD_STK` environment variable provides a non-intrusive way to request +# a larger pthread stack size for the tests. Various applications and runtime +# libraries on AIX use a default pthread stack size of 4 MiB, so we will use +# that as a default value here. +if "AIXTHREAD_STK" in os.environ: + config.environment["AIXTHREAD_STK"] = os.environ["AIXTHREAD_STK"] +elif platform.system() == "AIX": + config.environment["AIXTHREAD_STK"] = "4194304" + +# Some tools support an environment variable "OBJECT_MODE" on AIX OS, which +# controls the kind of objects they will support. If there is no "OBJECT_MODE" +# environment variable specified, the default behaviour is to support 32-bit +# objects only. In order to not affect most test cases, which expect to support +# 32-bit and 64-bit objects by default, set the environment variable +# "OBJECT_MODE" to "any" by default on AIX OS. + +if "system-aix" in config.available_features: + config.substitutions.append(("llvm-nm", "env OBJECT_MODE=any llvm-nm")) + config.substitutions.append(("llvm-ar", "env OBJECT_MODE=any llvm-ar")) + config.substitutions.append(("llvm-ranlib", "env OBJECT_MODE=any llvm-ranlib")) + +# It is not realistically possible to account for all options that could +# possibly be present in system and user configuration files, so disable +# default configs for the test runs. +config.environment["CLANG_NO_DEFAULT_CONFIG"] = "1" + +if lit_config.update_tests: + import sys + import os + + utilspath = os.path.join(config.llvm_src_root, "utils") + sys.path.append(utilspath) + from update_any_test_checks import utc_lit_plugin + + lit_config.test_updaters.append(utc_lit_plugin) >From d320ee632551a191a064e10514537e45295e1203 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 13:24:49 +0530 Subject: [PATCH 08/10] update lit.cfg.py --- clang/test/lit.cfg.py | 2 -- 1 file changed, 2 deletions(-) diff --git a/clang/test/lit.cfg.py b/clang/test/lit.cfg.py index 093400b0d708a..78f74039ff862 100644 --- a/clang/test/lit.cfg.py +++ b/clang/test/lit.cfg.py @@ -81,8 +81,6 @@ llvm_config.use_clang() -config.substitutions.append(('%clang_cir', os.path.join(config.clang_tools_dir, 'clang'))) - config.substitutions.append(("%src_dir", config.clang_src_dir)) config.substitutions.append(("%src_include_dir", config.clang_src_dir + "/include")) >From dd5108ba2c65d3cbd4506f5bffd964e7ee90cce7 Mon Sep 17 00:00:00 2001 From: Priyanshu3820 <[email protected]> Date: Thu, 11 Dec 2025 19:03:47 +0530 Subject: [PATCH 09/10] Update lit.cfg.py --- clang/test/lit.cfg.py | 888 +++++++++++++++++++++--------------------- 1 file changed, 444 insertions(+), 444 deletions(-) diff --git a/clang/test/lit.cfg.py b/clang/test/lit.cfg.py index 78f74039ff862..52b275c095475 100644 --- a/clang/test/lit.cfg.py +++ b/clang/test/lit.cfg.py @@ -1,444 +1,444 @@ -# -*- Python -*- - -import os -import platform -import re -import subprocess -import tempfile - -import lit.formats -import lit.util - -from lit.llvm import llvm_config -from lit.llvm.subst import ToolSubst -from lit.llvm.subst import FindTool - -# Configuration file for the 'lit' test runner. - -# name: The name of this test suite. -config.name = "Clang" - -# TODO: Consolidate the logic for turning on the internal shell by default for all LLVM test suites. -# See https://github.com/llvm/llvm-project/issues/106636 for more details. -# -# We prefer the lit internal shell which provides a better user experience on failures -# and is faster unless the user explicitly disables it with LIT_USE_INTERNAL_SHELL=0 -# env var. -use_lit_shell = True -lit_shell_env = os.environ.get("LIT_USE_INTERNAL_SHELL") -if lit_shell_env: - use_lit_shell = lit.util.pythonize_bool(lit_shell_env) - -# testFormat: The test format to use to interpret tests. -# -# For now we require '&&' between commands, until they get globally killed and -# the test runner updated. -config.test_format = lit.formats.ShTest(execute_external=not use_lit_shell) - -# suffixes: A list of file extensions to treat as test files. -config.suffixes = [ - ".c", - ".cpp", - ".i", - ".cir", - ".cppm", - ".m", - ".mm", - ".cu", - ".cuh", - ".hip", - ".hlsl", - ".ll", - ".cl", - ".clcpp", - ".s", - ".S", - ".modulemap", - ".test", - ".rs", - ".ifs", - ".rc", -] - -# excludes: A list of directories to exclude from the testsuite. The 'Inputs' -# subdirectories contain auxiliary inputs for various tests in their parent -# directories. -config.excludes = [ - "Inputs", - "CMakeLists.txt", - "README.txt", - "LICENSE.txt", - "debuginfo-tests", -] - -# test_source_root: The root path where tests are located. -config.test_source_root = os.path.dirname(__file__) - -# test_exec_root: The root path where tests should be run. -config.test_exec_root = os.path.join(config.clang_obj_root, "test") - -llvm_config.use_default_substitutions() - -llvm_config.use_clang() - -config.substitutions.append(("%src_dir", config.clang_src_dir)) - -config.substitutions.append(("%src_include_dir", config.clang_src_dir + "/include")) - -config.substitutions.append(("%target_triple", config.target_triple)) - -config.substitutions.append(("%PATH%", config.environment["PATH"])) - - -# For each occurrence of a clang tool name, replace it with the full path to -# the build directory holding that tool. We explicitly specify the directories -# to search to ensure that we get the tools just built and not some random -# tools that might happen to be in the user's PATH. -tool_dirs = [config.clang_tools_dir, config.llvm_tools_dir] - -tools = [ - "apinotes-test", - "c-index-test", - "cir-opt", - "clang-diff", - "clang-format", - "clang-repl", - "llvm-offload-binary", - "clang-tblgen", - "clang-scan-deps", - "clang-installapi", - "opt", - "llvm-ifs", - "yaml2obj", - "clang-linker-wrapper", - "clang-nvlink-wrapper", - "clang-sycl-linker", - "llvm-lto", - "llvm-lto2", - "llvm-profdata", - "llvm-readtapi", - ToolSubst( - "%clang_extdef_map", - command=FindTool("clang-extdef-mapping"), - unresolved="ignore", - ), -] - -if config.clang_examples: - config.available_features.add("examples") - - -def have_host_out_of_process_jit_feature_support(): - clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) - - if not clang_repl_exe: - return False - - testcode = b"\n".join([b"int i = 0;", b"%quit"]) - - try: - clang_repl_cmd = subprocess.run( - [clang_repl_exe, "-orc-runtime", "-oop-executor"], - stdout=subprocess.PIPE, - stderr=subprocess.PIPE, - input=testcode, - ) - except OSError: - return False - - if clang_repl_cmd.returncode == 0: - return True - - return False - - -def run_clang_repl(args): - clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) - - if not clang_repl_exe: - return "" - - try: - clang_repl_cmd = subprocess.Popen( - [clang_repl_exe, args], stdout=subprocess.PIPE - ) - except OSError: - print("could not exec clang-repl") - return "" - - clang_repl_out = clang_repl_cmd.stdout.read().decode("ascii") - clang_repl_cmd.wait() - - return clang_repl_out - - -def have_host_jit_feature_support(feature_name): - return "true" in run_clang_repl("--host-supports-" + feature_name) - -def have_host_clang_repl_cuda(): - clang_repl_exe = lit.util.which('clang-repl', config.clang_tools_dir) - - if not clang_repl_exe: - return False - - testcode = b'\n'.join([ - b"__global__ void test_func() {}", - b"test_func<<<1,1>>>();", - b"extern \"C\" int puts(const char *s);", - b"puts(cudaGetLastError() ? \"failure\" : \"success\");", - b"%quit" - ]) - try: - clang_repl_cmd = subprocess.run([clang_repl_exe, '--cuda'], - stdout=subprocess.PIPE, - stderr=subprocess.PIPE, - input=testcode) - except OSError: - return False - - if clang_repl_cmd.returncode == 0: - if clang_repl_cmd.stdout.find(b"success") != -1: - return True - - return False - -if have_host_jit_feature_support('jit'): - config.available_features.add('host-supports-jit') - - if have_host_clang_repl_cuda(): - config.available_features.add('host-supports-cuda') - hosttriple = run_clang_repl("--host-jit-triple") - config.substitutions.append(("%host-jit-triple", hosttriple.strip())) - - if have_host_out_of_process_jit_feature_support(): - config.available_features.add("host-supports-out-of-process-jit") - -if config.clang_staticanalyzer: - config.available_features.add("staticanalyzer") - tools.append("clang-check") - - if config.clang_staticanalyzer_z3: - config.available_features.add("z3") - if config.clang_staticanalyzer_z3_mock: - config.available_features.add("z3-mock") - else: - config.available_features.add("no-z3") - - check_analyzer_fixit_path = os.path.join( - config.test_source_root, "Analysis", "check-analyzer-fixit.py" - ) - config.substitutions.append( - ( - "%check_analyzer_fixit", - '"%s" %s' % (config.python_executable, check_analyzer_fixit_path), - ) - ) - - csv2json_path = os.path.join(config.test_source_root, "Analysis", "csv2json.py") - config.substitutions.append( - ( - "%csv2json", - '"%s" %s' % (config.python_executable, csv2json_path), - ) - ) - -# ClangIR support -if config.clang_enable_cir: - config.available_features.add("cir-support") - -llvm_config.add_tool_substitutions(tools, tool_dirs) - -config.substitutions.append( - ( - "%hmaptool", - "'%s' %s" - % ( - config.python_executable, - os.path.join(config.clang_src_dir, "utils", "hmaptool", "hmaptool"), - ), - ) -) - -config.substitutions.append( - ( - "%deps-to-rsp", - '"%s" %s' - % ( - config.python_executable, - os.path.join(config.clang_src_dir, "utils", "module-deps-to-rsp.py"), - ), - ) -) - -# Determine whether the test target is compatible with execution on the host. -if "aarch64" in config.host_arch: - config.available_features.add("aarch64-host") - -# Some tests are sensitive to whether clang is statically or dynamically linked -# to other libraries. -if not (config.build_shared_libs or config.link_llvm_dylib or config.link_clang_dylib): - config.available_features.add("static-libs") - -# Plugins (loadable modules) -if config.has_plugins and config.llvm_plugin_ext: - config.available_features.add("plugins") - -if config.clang_default_pie_on_linux: - config.available_features.add("default-pie-on-linux") - -# Set available features we allow tests to conditionalize on. -# -if config.clang_default_cxx_stdlib != "": - config.available_features.add( - "default-cxx-stdlib={}".format(config.clang_default_cxx_stdlib) - ) - -# As of 2011.08, crash-recovery tests still do not pass on FreeBSD. -if platform.system() not in ["FreeBSD"]: - config.available_features.add("crash-recovery") - -# ANSI escape sequences in non-dumb terminal -if platform.system() not in ["Windows"]: - config.available_features.add("ansi-escape-sequences") - -# Capability to print utf8 to the terminal. -# Windows expects codepage, unless Wide API. -if platform.system() not in ["Windows"]: - config.available_features.add("utf8-capable-terminal") - -# Support for libgcc runtime. Used to rule out tests that require -# clang to run with -rtlib=libgcc. -if platform.system() not in ["Darwin", "Fuchsia"]: - config.available_features.add("libgcc") - -# Case-insensitive file system - - -def is_filesystem_case_insensitive(): - os.makedirs(config.test_exec_root, exist_ok=True) - handle, path = tempfile.mkstemp(prefix="case-test", dir=config.test_exec_root) - isInsensitive = os.path.exists( - os.path.join(os.path.dirname(path), os.path.basename(path).upper()) - ) - os.close(handle) - os.remove(path) - return isInsensitive - - -if is_filesystem_case_insensitive(): - config.available_features.add("case-insensitive-filesystem") - -# Tests that require the /dev/fd filesystem. -if os.path.exists("/dev/fd/0") and sys.platform not in ["cygwin"]: - config.available_features.add("dev-fd-fs") - -# Set on native MS environment. -if re.match(r".*-(windows-msvc)$", config.target_triple): - config.available_features.add("ms-sdk") - -# [PR8833] LLP64-incompatible tests -if not re.match( - r"^(aarch64|arm64ec|x86_64).*-(windows-msvc|windows-gnu)$", config.target_triple -): - config.available_features.add("LP64") - -# Tests that are specific to the Apple Silicon macOS. -if re.match(r"^arm64(e)?-apple-(macos|darwin)", config.target_triple): - config.available_features.add("apple-silicon-mac") - -# [PR18856] Depends to remove opened file. On win32, a file could be removed -# only if all handles were closed. -if platform.system() not in ["Windows"]: - config.available_features.add("can-remove-opened-file") - -# Features -known_arches = ["x86_64", "mips64", "ppc64", "aarch64"] -if any(config.target_triple.startswith(x) for x in known_arches): - config.available_features.add("clang-target-64-bits") - - -def calculate_arch_features(arch_string): - features = [] - for arch in arch_string.split(): - features.append(arch.lower() + "-registered-target") - return features - - -llvm_config.feature_config( - [ - ("--assertion-mode", {"ON": "asserts"}), - ("--cxxflags", {r"-D_GLIBCXX_DEBUG\b": "libstdcxx-safe-mode"}), - ("--targets-built", calculate_arch_features), - ] -) - -if lit.util.which("xmllint"): - config.available_features.add("xmllint") - -if config.enable_backtrace: - config.available_features.add("backtrace") - -if config.enable_threads: - config.available_features.add("thread_support") - -# Check if we should allow outputs to console. -run_console_tests = int(lit_config.params.get("enable_console", "0")) -if run_console_tests != 0: - config.available_features.add("console") - -lit.util.usePlatformSdkOnDarwin(config, lit_config) -macOSSDKVersion = lit.util.findPlatformSdkVersionOnMacOS(config, lit_config) -if macOSSDKVersion is not None: - config.available_features.add("macos-sdk-" + str(macOSSDKVersion)) - -if os.path.exists("/etc/gentoo-release"): - config.available_features.add("gentoo") - -if config.enable_shared: - config.available_features.add("enable_shared") - -# Add a vendor-specific feature. -if config.clang_vendor_uti: - config.available_features.add("clang-vendor=" + config.clang_vendor_uti) - -if config.have_llvm_driver: - config.available_features.add("llvm-driver") - - -# Some tests perform deep recursion, which requires a larger pthread stack size -# than the relatively low default of 192 KiB for 64-bit processes on AIX. The -# `AIXTHREAD_STK` environment variable provides a non-intrusive way to request -# a larger pthread stack size for the tests. Various applications and runtime -# libraries on AIX use a default pthread stack size of 4 MiB, so we will use -# that as a default value here. -if "AIXTHREAD_STK" in os.environ: - config.environment["AIXTHREAD_STK"] = os.environ["AIXTHREAD_STK"] -elif platform.system() == "AIX": - config.environment["AIXTHREAD_STK"] = "4194304" - -# Some tools support an environment variable "OBJECT_MODE" on AIX OS, which -# controls the kind of objects they will support. If there is no "OBJECT_MODE" -# environment variable specified, the default behaviour is to support 32-bit -# objects only. In order to not affect most test cases, which expect to support -# 32-bit and 64-bit objects by default, set the environment variable -# "OBJECT_MODE" to "any" by default on AIX OS. - -if "system-aix" in config.available_features: - config.substitutions.append(("llvm-nm", "env OBJECT_MODE=any llvm-nm")) - config.substitutions.append(("llvm-ar", "env OBJECT_MODE=any llvm-ar")) - config.substitutions.append(("llvm-ranlib", "env OBJECT_MODE=any llvm-ranlib")) - -# It is not realistically possible to account for all options that could -# possibly be present in system and user configuration files, so disable -# default configs for the test runs. -config.environment["CLANG_NO_DEFAULT_CONFIG"] = "1" - -if lit_config.update_tests: - import sys - import os - - utilspath = os.path.join(config.llvm_src_root, "utils") - sys.path.append(utilspath) - from update_any_test_checks import utc_lit_plugin - - lit_config.test_updaters.append(utc_lit_plugin) +# -*- Python -*- + +import os +import platform +import re +import subprocess +import tempfile + +import lit.formats +import lit.util + +from lit.llvm import llvm_config +from lit.llvm.subst import ToolSubst +from lit.llvm.subst import FindTool + +# Configuration file for the 'lit' test runner. + +# name: The name of this test suite. +config.name = "Clang" + +# TODO: Consolidate the logic for turning on the internal shell by default for all LLVM test suites. +# See https://github.com/llvm/llvm-project/issues/106636 for more details. +# +# We prefer the lit internal shell which provides a better user experience on failures +# and is faster unless the user explicitly disables it with LIT_USE_INTERNAL_SHELL=0 +# env var. +use_lit_shell = True +lit_shell_env = os.environ.get("LIT_USE_INTERNAL_SHELL") +if lit_shell_env: + use_lit_shell = lit.util.pythonize_bool(lit_shell_env) + +# testFormat: The test format to use to interpret tests. +# +# For now we require '&&' between commands, until they get globally killed and +# the test runner updated. +config.test_format = lit.formats.ShTest(execute_external=not use_lit_shell) + +# suffixes: A list of file extensions to treat as test files. +config.suffixes = [ + ".c", + ".cpp", + ".i", + ".cir", + ".cppm", + ".m", + ".mm", + ".cu", + ".cuh", + ".hip", + ".hlsl", + ".ll", + ".cl", + ".clcpp", + ".s", + ".S", + ".modulemap", + ".test", + ".rs", + ".ifs", + ".rc", +] + +# excludes: A list of directories to exclude from the testsuite. The 'Inputs' +# subdirectories contain auxiliary inputs for various tests in their parent +# directories. +config.excludes = [ + "Inputs", + "CMakeLists.txt", + "README.txt", + "LICENSE.txt", + "debuginfo-tests", +] + +# test_source_root: The root path where tests are located. +config.test_source_root = os.path.dirname(__file__) + +# test_exec_root: The root path where tests should be run. +config.test_exec_root = os.path.join(config.clang_obj_root, "test") + +llvm_config.use_default_substitutions() + +llvm_config.use_clang() + +config.substitutions.append(("%src_dir", config.clang_src_dir)) + +config.substitutions.append(("%src_include_dir", config.clang_src_dir + "/include")) + +config.substitutions.append(("%target_triple", config.target_triple)) + +config.substitutions.append(("%PATH%", config.environment["PATH"])) + + +# For each occurrence of a clang tool name, replace it with the full path to +# the build directory holding that tool. We explicitly specify the directories +# to search to ensure that we get the tools just built and not some random +# tools that might happen to be in the user's PATH. +tool_dirs = [config.clang_tools_dir, config.llvm_tools_dir] + +tools = [ + "apinotes-test", + "c-index-test", + "cir-opt", + "clang-diff", + "clang-format", + "clang-repl", + "llvm-offload-binary", + "clang-tblgen", + "clang-scan-deps", + "clang-installapi", + "opt", + "llvm-ifs", + "yaml2obj", + "clang-linker-wrapper", + "clang-nvlink-wrapper", + "clang-sycl-linker", + "llvm-lto", + "llvm-lto2", + "llvm-profdata", + "llvm-readtapi", + ToolSubst( + "%clang_extdef_map", + command=FindTool("clang-extdef-mapping"), + unresolved="ignore", + ), +] + +if config.clang_examples: + config.available_features.add("examples") + + +def have_host_out_of_process_jit_feature_support(): + clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) + + if not clang_repl_exe: + return False + + testcode = b"\n".join([b"int i = 0;", b"%quit"]) + + try: + clang_repl_cmd = subprocess.run( + [clang_repl_exe, "-orc-runtime", "-oop-executor"], + stdout=subprocess.PIPE, + stderr=subprocess.PIPE, + input=testcode, + ) + except OSError: + return False + + if clang_repl_cmd.returncode == 0: + return True + + return False + + +def run_clang_repl(args): + clang_repl_exe = lit.util.which("clang-repl", config.clang_tools_dir) + + if not clang_repl_exe: + return "" + + try: + clang_repl_cmd = subprocess.Popen( + [clang_repl_exe, args], stdout=subprocess.PIPE + ) + except OSError: + print("could not exec clang-repl") + return "" + + clang_repl_out = clang_repl_cmd.stdout.read().decode("ascii") + clang_repl_cmd.wait() + + return clang_repl_out + + +def have_host_jit_feature_support(feature_name): + return "true" in run_clang_repl("--host-supports-" + feature_name) + +def have_host_clang_repl_cuda(): + clang_repl_exe = lit.util.which('clang-repl', config.clang_tools_dir) + + if not clang_repl_exe: + return False + + testcode = b'\n'.join([ + b"__global__ void test_func() {}", + b"test_func<<<1,1>>>();", + b"extern \"C\" int puts(const char *s);", + b"puts(cudaGetLastError() ? \"failure\" : \"success\");", + b"%quit" + ]) + try: + clang_repl_cmd = subprocess.run([clang_repl_exe, '--cuda'], + stdout=subprocess.PIPE, + stderr=subprocess.PIPE, + input=testcode) + except OSError: + return False + + if clang_repl_cmd.returncode == 0: + if clang_repl_cmd.stdout.find(b"success") != -1: + return True + + return False + +if have_host_jit_feature_support('jit'): + config.available_features.add('host-supports-jit') + + if have_host_clang_repl_cuda(): + config.available_features.add('host-supports-cuda') + hosttriple = run_clang_repl("--host-jit-triple") + config.substitutions.append(("%host-jit-triple", hosttriple.strip())) + + if have_host_out_of_process_jit_feature_support(): + config.available_features.add("host-supports-out-of-process-jit") + +if config.clang_staticanalyzer: + config.available_features.add("staticanalyzer") + tools.append("clang-check") + + if config.clang_staticanalyzer_z3: + config.available_features.add("z3") + if config.clang_staticanalyzer_z3_mock: + config.available_features.add("z3-mock") + else: + config.available_features.add("no-z3") + + check_analyzer_fixit_path = os.path.join( + config.test_source_root, "Analysis", "check-analyzer-fixit.py" + ) + config.substitutions.append( + ( + "%check_analyzer_fixit", + '"%s" %s' % (config.python_executable, check_analyzer_fixit_path), + ) + ) + + csv2json_path = os.path.join(config.test_source_root, "Analysis", "csv2json.py") + config.substitutions.append( + ( + "%csv2json", + '"%s" %s' % (config.python_executable, csv2json_path), + ) + ) + +# ClangIR support +if config.clang_enable_cir: + config.available_features.add("cir-support") + +llvm_config.add_tool_substitutions(tools, tool_dirs) + +config.substitutions.append( + ( + "%hmaptool", + "'%s' %s" + % ( + config.python_executable, + os.path.join(config.clang_src_dir, "utils", "hmaptool", "hmaptool"), + ), + ) +) + +config.substitutions.append( + ( + "%deps-to-rsp", + '"%s" %s' + % ( + config.python_executable, + os.path.join(config.clang_src_dir, "utils", "module-deps-to-rsp.py"), + ), + ) +) + +# Determine whether the test target is compatible with execution on the host. +if "aarch64" in config.host_arch: + config.available_features.add("aarch64-host") + +# Some tests are sensitive to whether clang is statically or dynamically linked +# to other libraries. +if not (config.build_shared_libs or config.link_llvm_dylib or config.link_clang_dylib): + config.available_features.add("static-libs") + +# Plugins (loadable modules) +if config.has_plugins and config.llvm_plugin_ext: + config.available_features.add("plugins") + +if config.clang_default_pie_on_linux: + config.available_features.add("default-pie-on-linux") + +# Set available features we allow tests to conditionalize on. +# +if config.clang_default_cxx_stdlib != "": + config.available_features.add( + "default-cxx-stdlib={}".format(config.clang_default_cxx_stdlib) + ) + +# As of 2011.08, crash-recovery tests still do not pass on FreeBSD. +if platform.system() not in ["FreeBSD"]: + config.available_features.add("crash-recovery") + +# ANSI escape sequences in non-dumb terminal +if platform.system() not in ["Windows"]: + config.available_features.add("ansi-escape-sequences") + +# Capability to print utf8 to the terminal. +# Windows expects codepage, unless Wide API. +if platform.system() not in ["Windows"]: + config.available_features.add("utf8-capable-terminal") + +# Support for libgcc runtime. Used to rule out tests that require +# clang to run with -rtlib=libgcc. +if platform.system() not in ["Darwin", "Fuchsia"]: + config.available_features.add("libgcc") + +# Case-insensitive file system + + +def is_filesystem_case_insensitive(): + os.makedirs(config.test_exec_root, exist_ok=True) + handle, path = tempfile.mkstemp(prefix="case-test", dir=config.test_exec_root) + isInsensitive = os.path.exists( + os.path.join(os.path.dirname(path), os.path.basename(path).upper()) + ) + os.close(handle) + os.remove(path) + return isInsensitive + + +if is_filesystem_case_insensitive(): + config.available_features.add("case-insensitive-filesystem") + +# Tests that require the /dev/fd filesystem. +if os.path.exists("/dev/fd/0") and sys.platform not in ["cygwin"]: + config.available_features.add("dev-fd-fs") + +# Set on native MS environment. +if re.match(r".*-(windows-msvc)$", config.target_triple): + config.available_features.add("ms-sdk") + +# [PR8833] LLP64-incompatible tests +if not re.match( + r"^(aarch64|arm64ec|x86_64).*-(windows-msvc|windows-gnu)$", config.target_triple +): + config.available_features.add("LP64") + +# Tests that are specific to the Apple Silicon macOS. +if re.match(r"^arm64(e)?-apple-(macos|darwin)", config.target_triple): + config.available_features.add("apple-silicon-mac") + +# [PR18856] Depends to remove opened file. On win32, a file could be removed +# only if all handles were closed. +if platform.system() not in ["Windows"]: + config.available_features.add("can-remove-opened-file") + +# Features +known_arches = ["x86_64", "mips64", "ppc64", "aarch64"] +if any(config.target_triple.startswith(x) for x in known_arches): + config.available_features.add("clang-target-64-bits") + + +def calculate_arch_features(arch_string): + features = [] + for arch in arch_string.split(): + features.append(arch.lower() + "-registered-target") + return features + + +llvm_config.feature_config( + [ + ("--assertion-mode", {"ON": "asserts"}), + ("--cxxflags", {r"-D_GLIBCXX_DEBUG\b": "libstdcxx-safe-mode"}), + ("--targets-built", calculate_arch_features), + ] +) + +if lit.util.which("xmllint"): + config.available_features.add("xmllint") + +if config.enable_backtrace: + config.available_features.add("backtrace") + +if config.enable_threads: + config.available_features.add("thread_support") + +# Check if we should allow outputs to console. +run_console_tests = int(lit_config.params.get("enable_console", "0")) +if run_console_tests != 0: + config.available_features.add("console") + +lit.util.usePlatformSdkOnDarwin(config, lit_config) +macOSSDKVersion = lit.util.findPlatformSdkVersionOnMacOS(config, lit_config) +if macOSSDKVersion is not None: + config.available_features.add("macos-sdk-" + str(macOSSDKVersion)) + +if os.path.exists("/etc/gentoo-release"): + config.available_features.add("gentoo") + +if config.enable_shared: + config.available_features.add("enable_shared") + +# Add a vendor-specific feature. +if config.clang_vendor_uti: + config.available_features.add("clang-vendor=" + config.clang_vendor_uti) + +if config.have_llvm_driver: + config.available_features.add("llvm-driver") + + +# Some tests perform deep recursion, which requires a larger pthread stack size +# than the relatively low default of 192 KiB for 64-bit processes on AIX. The +# `AIXTHREAD_STK` environment variable provides a non-intrusive way to request +# a larger pthread stack size for the tests. Various applications and runtime +# libraries on AIX use a default pthread stack size of 4 MiB, so we will use +# that as a default value here. +if "AIXTHREAD_STK" in os.environ: + config.environment["AIXTHREAD_STK"] = os.environ["AIXTHREAD_STK"] +elif platform.system() == "AIX": + config.environment["AIXTHREAD_STK"] = "4194304" + +# Some tools support an environment variable "OBJECT_MODE" on AIX OS, which +# controls the kind of objects they will support. If there is no "OBJECT_MODE" +# environment variable specified, the default behaviour is to support 32-bit +# objects only. In order to not affect most test cases, which expect to support +# 32-bit and 64-bit objects by default, set the environment variable +# "OBJECT_MODE" to "any" by default on AIX OS. + +if "system-aix" in config.available_features: + config.substitutions.append(("llvm-nm", "env OBJECT_MODE=any llvm-nm")) + config.substitutions.append(("llvm-ar", "env OBJECT_MODE=any llvm-ar")) + config.substitutions.append(("llvm-ranlib", "env OBJECT_MODE=any llvm-ranlib")) + +# It is not realistically possible to account for all options that could +# possibly be present in system and user configuration files, so disable +# default configs for the test runs. +config.environment["CLANG_NO_DEFAULT_CONFIG"] = "1" + +if lit_config.update_tests: + import sys + import os + + utilspath = os.path.join(config.llvm_src_root, "utils") + sys.path.append(utilspath) + from update_any_test_checks import utc_lit_plugin + + lit_config.test_updaters.append(utc_lit_plugin) >From 608ee13a646dacbb83393f8d75611c8105289432 Mon Sep 17 00:00:00 2001 From: Priyanshu Kumar <[email protected]> Date: Thu, 11 Dec 2025 19:27:20 +0000 Subject: [PATCH 10/10] Update test --- .../CodeGenBuiltins/X86/cir-convert-half.c | 99 ++++++++----------- 1 file changed, 41 insertions(+), 58 deletions(-) diff --git a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c index 09c08ef06a133..29ed56e154a7b 100644 --- a/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c +++ b/clang/test/CIR/CodeGenBuiltins/X86/cir-convert-half.c @@ -1,84 +1,67 @@ -// Test X86-specific convert_half builtins +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-cir %s -o - | FileCheck %s --check-prefix=CIR +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o - | FileCheck %s --check-prefix=LLVM +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o - | FileCheck %s --check-prefix=OGCG -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-cir %s -o %t.cir -// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -fclangir -emit-llvm %s -o %t-cir.ll -// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s -// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -target-feature +avx512fp16 -target-feature +avx512bf16 -emit-llvm %s -o %t.ll -// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s +// REQUIRES: avx512fp16 +// REQUIRES: avx512bf16 + +#include <stdint.h> +#ifndef __bf16 +typedef uint16_t __bf16; +#endif typedef float __m512 __attribute__((__vector_size__(64), __aligned__(64))); typedef float __m256 __attribute__((__vector_size__(32), __aligned__(32))); typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16))); -typedef int __m256i __attribute__((__vector_size__(32), __aligned__(32))); -typedef int __m128i __attribute__((__vector_size__(16), __aligned__(16))); typedef int __mmask16; -typedef unsigned char __mmask8; -typedef __bf16 __m256bh __attribute__((__vector_size__(32), __aligned__(32))); -typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16))); +typedef int __mmask8; +typedef short __m256i __attribute__((__vector_size__(32), __aligned__(32))); +typedef short __m128i __attribute__((__vector_size__(16), __aligned__(16))); -// Test __builtin_ia32_vcvtph2ps512_mask +// CIR-LABEL: test_vcvtph2ps512_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.512" +// LLVM-LABEL: test_vcvtph2ps512_mask +// OGCG-LABEL: test_vcvtph2ps512_mask __m512 test_vcvtph2ps512_mask(__m256i a, __m512 src, __mmask16 k) { return __builtin_ia32_vcvtph2ps512_mask(a, src, k, 4); } -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps512_mask -// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.512" -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps512_mask -// LLVM: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps512_mask -// OGCG: call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512 -// Test __builtin_ia32_vcvtph2ps256_mask +// CIR-LABEL: test_vcvtph2ps256_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.256" +// LLVM-LABEL: test_vcvtph2ps256_mask +// OGCG-LABEL: test_vcvtph2ps256_mask __m256 test_vcvtph2ps256_mask(__m128i a, __m256 src, __mmask8 k) { return __builtin_ia32_vcvtph2ps256_mask(a, src, k); } -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps256_mask -// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.256" -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps256_mask -// LLVM: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps256_mask -// OGCG: call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256 -// Test __builtin_ia32_vcvtph2ps_mask +// CIR-LABEL: test_vcvtph2ps_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.128" +// LLVM-LABEL: test_vcvtph2ps_mask +// OGCG-LABEL: test_vcvtph2ps_mask __m128 test_vcvtph2ps_mask(__m128i a, __m128 src, __mmask8 k) { return __builtin_ia32_vcvtph2ps_mask(a, src, k); } -// CIR-LABEL: cir.func {{.*}}@test_vcvtph2ps_mask -// CIR: cir.call_llvm_intrinsic "x86.avx512.mask.vcvtph2ps.128" -// LLVM-LABEL: define {{.*}} @test_vcvtph2ps_mask -// LLVM: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 -// OGCG-LABEL: define {{.*}} @test_vcvtph2ps_mask -// OGCG: call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128 -// Test __builtin_ia32_cvtneps2bf16_512_mask -__m256bh test_cvtneps2bf16_512_mask(__m512 a, __m256bh w, __mmask16 u) { +// CIR-LABEL: test_cvtneps2bf16_512_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.512" +// LLVM-LABEL: test_cvtneps2bf16_512_mask +// OGCG-LABEL: test_cvtneps2bf16_512_mask +__bf16 test_cvtneps2bf16_512_mask(__m512 a, __bf16 w, __mmask16 u) { return __builtin_ia32_cvtneps2bf16_512_mask(a, w, u); } -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_512_mask -// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.512" -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask -// LLVM: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_512_mask -// OGCG: call <32 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.512 -// Test __builtin_ia32_cvtneps2bf16_256_mask -__m128bh test_cvtneps2bf16_256_mask(__m256 a, __m128bh w, __mmask8 u) { +// CIR-LABEL: test_cvtneps2bf16_256_mask +// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.256" +// LLVM-LABEL: test_cvtneps2bf16_256_mask +// OGCG-LABEL: test_cvtneps2bf16_256_mask +__bf16 test_cvtneps2bf16_256_mask(__m256 a, __bf16 w, __mmask8 u) { return __builtin_ia32_cvtneps2bf16_256_mask(a, w, u); } -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_256_mask -// CIR: cir.call_llvm_intrinsic "x86.avx512bf16.cvtneps2bf16.256" -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask -// LLVM: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_256_mask -// OGCG: call <16 x bfloat> @llvm.x86.avx512bf16.cvtneps2bf16.256 -// Test __builtin_ia32_cvtneps2bf16_128_mask -__m128bh test_cvtneps2bf16_128_mask(__m128 a, __m128bh w, __mmask8 u) { - return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); -} -// CIR-LABEL: cir.func {{.*}}@test_cvtneps2bf16_128_mask +// CIR-LABEL: test_cvtneps2bf16_128_mask // CIR: cir.call_llvm_intrinsic "x86.avx512bf16.mask.cvtneps2bf16.128" -// LLVM-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask -// LLVM: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 -// OGCG-LABEL: define {{.*}} @test_cvtneps2bf16_128_mask -// OGCG: call <8 x bfloat> @llvm.x86.avx512bf16.mask.cvtneps2bf16.128 \ No newline at end of file +// LLVM-LABEL: test_cvtneps2bf16_128_mask +// OGCG-LABEL: test_cvtneps2bf16_128_mask +__bf16 test_cvtneps2bf16_128_mask(__m128 a, __bf16 w, __mmask8 u) { + return __builtin_ia32_cvtneps2bf16_128_mask(a, w, u); +} \ No newline at end of file _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
