diff options
Diffstat (limited to 'clang/test/CIR/CodeGen/complex-compound-assignment.cpp')
| -rw-r--r-- | clang/test/CIR/CodeGen/complex-compound-assignment.cpp | 290 |
1 files changed, 288 insertions, 2 deletions
diff --git a/clang/test/CIR/CodeGen/complex-compound-assignment.cpp b/clang/test/CIR/CodeGen/complex-compound-assignment.cpp index 9a6659bd5d93..9909985e7819 100644 --- a/clang/test/CIR/CodeGen/complex-compound-assignment.cpp +++ b/clang/test/CIR/CodeGen/complex-compound-assignment.cpp @@ -198,9 +198,9 @@ void foo3() { // LLVM: %[[RESULT_REAL:.*]] = extractvalue { float, float } %[[RESULT]], 0 // LLVM: %[[RESULT_IMAG:.*]] = extractvalue { float, float } %[[RESULT]], 1 // LLVM: %[[RESULT_REAL_F16:.*]] = fptrunc float %[[RESULT_REAL]] to half -// LLVM: %[[RESULT_IMAG_F26:.*]] = fptrunc float %[[RESULT_IMAG]] to half +// LLVM: %[[RESULT_IMAG_F16:.*]] = fptrunc float %[[RESULT_IMAG]] to half // LLVM: %[[TMP_RESULT_F16:.*]] = insertvalue { half, half } undef, half %[[RESULT_REAL_F16]], 0 -// LLVM: %[[RESULT_F16:.*]] = insertvalue { half, half } %29, half %[[RESULT_IMAG_F26]], 1 +// LLVM: %[[RESULT_F16:.*]] = insertvalue { half, half } %29, half %[[RESULT_IMAG_F16]], 1 // LLVM: store { half, half } %[[RESULT_F16]], ptr %[[B_ADDR]], align 2 // OGCG: %[[A_ADDR:.*]] = alloca { half, half }, align 2 @@ -534,6 +534,292 @@ void foo8() { // OGCG: store float %[[RESULT_REAL]], ptr %[[A_REAL_PTR]], align 4 // OGCG: store float %[[RESULT_IMAG]], ptr %[[A_IMAG_PTR]], align 4 +void foo10() { + float _Complex a; + float _Complex b; + a /= b; +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>, ["a"] +// CIR: %[[B_ADDR:.*]] = cir.alloca !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>, ["b"] +// CIR: %[[TMP_B:.*]] = cir.load{{.*}} %[[B_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, !cir.complex<!cir.float> +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, !cir.complex<!cir.float> +// CIR: %[[A_REAL:.*]] = cir.complex.real %[[TMP_A]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[A_IMAG:.*]] = cir.complex.imag %[[TMP_A]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[B_REAL:.*]] = cir.complex.real %[[TMP_B]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[B_IMAG:.*]] = cir.complex.imag %[[TMP_B]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[RESULT:.*]] = cir.call @__divsc3(%[[A_REAL]], %[[A_IMAG]], %[[B_REAL]], %[[B_IMAG]]) : (!cir.float, !cir.float, !cir.float, !cir.float) -> !cir.complex<!cir.float> +// CIR: cir.store{{.*}} %[[RESULT]], %[[A_ADDR]] : !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>> + +// LLVM: %[[A_ADDR:.*]] = alloca { float, float }, i64 1, align 4 +// LLVM: %[[B_ADDR:.*]] = alloca { float, float }, i64 1, align 4 +// LLVM: %[[TMP_B:.*]] = load { float, float }, ptr %[[B_ADDR]], align 4 +// LLVM: %[[TMP_A:.*]] = load { float, float }, ptr %[[A_ADDR]], align 4 +// LLVM: %[[A_REAL:.*]] = extractvalue { float, float } %[[TMP_A]], 0 +// LLVM: %[[A_IMAG:.*]] = extractvalue { float, float } %[[TMP_A]], 1 +// LLVM: %[[B_REAL:.*]] = extractvalue { float, float } %[[TMP_B]], 0 +// LLVM: %[[B_IMAG:.*]] = extractvalue { float, float } %[[TMP_B]], 1 +// LLVM: %[[RESULT:.*]] = call { float, float } @__divsc3(float %[[A_REAL]], float %[[A_IMAG]], float %[[B_REAL]], float %[[B_IMAG]]) +// LLVM: store { float, float } %[[RESULT]], ptr %[[A_ADDR]], align 4 + +// OGCG: %[[A_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[B_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[RESULT_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[B_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[B_ADDR]], i32 0, i32 0 +// OGCG: %[[B_REAL:.*]] = load float, ptr %[[B_REAL_PTR]], align 4 +// OGCG: %[[B_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[B_ADDR]], i32 0, i32 1 +// OGCG: %[[B_IMAG:.*]] = load float, ptr %[[B_IMAG_PTR]], align 4 +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_REAL:.*]] = load float, ptr %[[A_REAL_PTR]], align 4 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: %[[A_IMAG:.*]] = load float, ptr %[[A_IMAG_PTR]], align 4 +// OGCG: %[[RESULT:.*]] = call{{.*}} <2 x float> @__divsc3(float noundef %[[A_REAL]], float noundef %[[A_IMAG]], float noundef %[[B_REAL]], float noundef %[[B_IMAG]]) +// OGCG: store <2 x float> %[[RESULT]], ptr %[[RESULT_ADDR]], align 4 +// OGCG: %[[RESULT_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RESULT_ADDR]], i32 0, i32 0 +// OGCG: %[[RESULT_REAL:.*]] = load float, ptr %[[RESULT_REAL_PTR]], align 4 +// OGCG: %[[RESULT_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RESULT_ADDR]], i32 0, i32 1 +// OGCG: %[[RESULT_IMAG:.*]] = load float, ptr %[[RESULT_IMAG_PTR]], align 4 +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: store float %[[RESULT_REAL]], ptr %[[A_REAL_PTR]], align 4 +// OGCG: store float %[[RESULT_IMAG]], ptr %[[A_IMAG_PTR]], align 4 + +void foo11() { + float _Complex a; + float b; + a /= b; +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>, ["a"] +// CIR: %[[B_ADDR:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["b"] +// CIR: %[[TMP_B:.*]] = cir.load{{.*}} %[[B_ADDR]] : !cir.ptr<!cir.float>, !cir.float +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, !cir.complex<!cir.float> +// CIR: %[[A_REAL:.*]] = cir.complex.real %[[TMP_A]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[A_IMAG:.*]] = cir.complex.imag %[[TMP_A]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[RESULT_REAL:.*]] = cir.binop(div, %[[A_REAL]], %[[TMP_B]]) : !cir.float +// CIR: %[[RESULT_IMAG:.*]] = cir.binop(div, %[[A_IMAG]], %[[TMP_B]]) : !cir.float +// CIR: %[[RESULT:.*]] = cir.complex.create %[[RESULT_REAL]], %[[RESULT_IMAG]] : !cir.float -> !cir.complex<!cir.float> +// CIR: cir.store{{.*}} %[[RESULT]], %[[A_ADDR]] : !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>> + +// LLVM: %[[A_ADDR:.*]] = alloca { float, float }, i64 1, align 4 +// LLVM: %[[B_ADDR:.*]] = alloca float, i64 1, align 4 +// LLVM: %[[TMP_B:.*]] = load float, ptr %[[B_ADDR]], align 4 +// LLVM: %[[TMP_A:.*]] = load { float, float }, ptr %[[A_ADDR]], align 4 +// LLVM: %[[A_REAL:.*]] = extractvalue { float, float } %[[TMP_A]], 0 +// LLVM: %[[A_IMAG:.*]] = extractvalue { float, float } %[[TMP_A]], 1 +// LLVM: %[[RESULT_REAL:.*]] = fdiv float %[[A_REAL]], %[[TMP_B]] +// LLVM: %[[RESULT_IMAG:.*]] = fdiv float %[[A_IMAG]], %[[TMP_B]] +// LLVM: %[[TMP_RESULT:.*]] = insertvalue { float, float } {{.*}}, float %[[RESULT_REAL]], 0 +// LLVM: %[[RESULT:.*]] = insertvalue { float, float } %[[TMP_RESULT]], float %[[RESULT_IMAG]], 1 +// LLVM: store { float, float } %[[RESULT]], ptr %[[A_ADDR]], align 4 + +// OGCG: %[[A_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[B_ADDR:.*]] = alloca float, align 4 +// OGCG: %[[TMP_B:.*]] = load float, ptr %[[B_ADDR]], align 4 +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_REAL:.*]] = load float, ptr %[[A_REAL_PTR]], align 4 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: %[[A_IMAG:.*]] = load float, ptr %[[A_IMAG_PTR]], align 4 +// OGCG: %[[RESULT_REAL:.*]] = fdiv float %[[A_REAL]], %[[TMP_B]] +// OGCG: %[[RESULT_IMAG:.*]] = fdiv float %[[A_IMAG]], %[[TMP_B]] +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: store float %[[RESULT_REAL]], ptr %[[A_REAL_PTR]], align 4 +// OGCG: store float %[[RESULT_IMAG]], ptr %[[A_IMAG_PTR]], align 4 + +void foo12() { + int _Complex a; + int b; + a /= b; +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.complex<!s32i>, !cir.ptr<!cir.complex<!s32i>>, ["a"] +// CIR: %[[B_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["b"] +// CIR: %[[TMP_B:.*]] = cir.load{{.*}} %[[B_ADDR]] : !cir.ptr<!s32i>, !s32i +// CIR: %[[CONST_0:.*]] = cir.const #cir.int<0> : !s32i +// CIR: %[[B_COMPLEX:.*]] = cir.complex.create %[[TMP_B]], %[[CONST_0]] : !s32i -> !cir.complex<!s32i> +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.complex<!s32i>>, !cir.complex<!s32i> +// CIR: %[[A_REAL:.*]] = cir.complex.real %[[TMP_A]] : !cir.complex<!s32i> -> !s32i +// CIR: %[[A_IMAG:.*]] = cir.complex.imag %[[TMP_A]] : !cir.complex<!s32i> -> !s32i +// CIR: %[[B_REAL:.*]] = cir.complex.real %[[B_COMPLEX]] : !cir.complex<!s32i> -> !s32i +// CIR: %[[B_IMAG:.*]] = cir.complex.imag %[[B_COMPLEX]] : !cir.complex<!s32i> -> !s32i +// CIR: %[[MUL_AR_BR:.*]] = cir.binop(mul, %[[A_REAL]], %[[B_REAL]]) : !s32i +// CIR: %[[MUL_AI_BI:.*]] = cir.binop(mul, %[[A_IMAG]], %[[B_IMAG]]) : !s32i +// CIR: %[[MUL_BR_BR:.*]] = cir.binop(mul, %[[B_REAL]], %[[B_REAL]]) : !s32i +// CIR: %[[MUL_BI_BI:.*]] = cir.binop(mul, %[[B_IMAG]], %[[B_IMAG]]) : !s32i +// CIR: %[[ADD_ARBR_AIBI:.*]] = cir.binop(add, %[[MUL_AR_BR]], %[[MUL_AI_BI]]) : !s32i +// CIR: %[[ADD_BRBR_BIBI:.*]] = cir.binop(add, %[[MUL_BR_BR]], %[[MUL_BI_BI]]) : !s32i +// CIR: %[[RESULT_REAL:.*]] = cir.binop(div, %[[ADD_ARBR_AIBI]], %[[ADD_BRBR_BIBI]]) : !s32i +// CIR: %[[MUL_AI_BR:.*]] = cir.binop(mul, %[[A_IMAG]], %[[B_REAL]]) : !s32i +// CIR: %[[MUL_AR_BI:.*]] = cir.binop(mul, %[[A_REAL]], %[[B_IMAG]]) : !s32i +// CIR: %[[SUB_AIBR_ARBI:.*]] = cir.binop(sub, %[[MUL_AI_BR]], %[[MUL_AR_BI]]) : !s32i +// CIR: %[[RESULT_IMAG:.*]] = cir.binop(div, %[[SUB_AIBR_ARBI]], %[[ADD_BRBR_BIBI]]) : !s32i +// CIR: %[[RESULT:.*]] = cir.complex.create %[[RESULT_REAL]], %[[RESULT_IMAG]] : !s32i -> !cir.complex<!s32i> +// CIR: cir.store{{.*}} %[[RESULT]], %[[A_ADDR]] : !cir.complex<!s32i>, !cir.ptr<!cir.complex<!s32i>> + +// LLVM: %[[A_ADDR:.*]] = alloca { i32, i32 }, i64 1, align 4 +// LLVM: %[[B_ADDR:.*]] = alloca i32, i64 1, align 4 +// LLVM: %[[TMP_B:.*]] = load i32, ptr %[[B_ADDR]], align 4 +// LLVM: %[[TMP_B_COMPLEX:.*]] = insertvalue { i32, i32 } {{.*}}, i32 %[[TMP_B]], 0 +// LLVM: %[[B_COMPLEX:.*]] = insertvalue { i32, i32 } %[[TMP_B_COMPLEX]], i32 0, 1 +// LLVM: %[[TMP_A:.*]] = load { i32, i32 }, ptr %[[A_ADDR]], align 4 +// LLVM: %[[A_REAL:.*]] = extractvalue { i32, i32 } %[[TMP_A]], 0 +// LLVM: %[[A_IMAG:.*]] = extractvalue { i32, i32 } %[[TMP_A]], 1 +// LLVM: %[[MUL_AR_BR:.*]] = mul i32 %[[A_REAL]], %[[TMP_B]] +// LLVM: %[[MUL_AI_BI:.*]] = mul i32 %[[A_IMAG]], 0 +// LLVM: %[[MUL_BR_BR:.*]] = mul i32 %[[TMP_B]], %[[TMP_B]] +// LLVM: %[[ADD_ARBR_AIBI:.*]] = add i32 %[[MUL_AR_BR]], %[[MUL_AI_BI]] +// LLVM: %[[ADD_BRBR_BIBI:.*]] = add i32 %[[MUL_BR_BR]], 0 +// LLVM: %[[RESULT_REAL:.*]] = sdiv i32 %[[ADD_ARBR_AIBI]], %[[ADD_BRBR_BIBI]] +// LLVM: %[[MUL_AI_BR:.*]] = mul i32 %[[A_IMAG]], %[[TMP_B]] +// LLVM: %[[MUL_AR_BI:.*]] = mul i32 %[[A_REAL]], 0 +// LLVM: %[[SUB_AIBR_ARBI:.*]] = sub i32 %[[MUL_AI_BR]], %[[MUL_AR_BI]] +// LLVM: %[[RESULT_IMAG:.*]] = sdiv i32 %[[SUB_AIBR_ARBI]], %[[ADD_BRBR_BIBI]] +// LLVM: %[[TMP_RESULT:.*]] = insertvalue { i32, i32 } {{.*}}, i32 %[[RESULT_REAL]], 0 +// LLVM: %[[RESULT:.*]] = insertvalue { i32, i32 } %[[TMP_RESULT]], i32 %[[RESULT_IMAG]], 1 +// LLVM: store { i32, i32 } %[[RESULT]], ptr %[[A_ADDR]], align 4 + +// OGCG: %[[A_ADDR:.*]] = alloca { i32, i32 }, align 4 +// OGCG: %[[B_ADDR:.*]] = alloca i32, align 4 +// OGCG: %[[TMP_B:.*]] = load i32, ptr %[[B_ADDR]], align 4 +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { i32, i32 }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_REAL:.*]] = load i32, ptr %[[A_REAL_PTR]], align 4 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { i32, i32 }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: %[[A_IMAG:.*]] = load i32, ptr %[[A_IMAG_PTR]], align 4 +// OGCG: %[[MUL_AR_BR:.*]] = mul i32 %[[A_REAL]], %[[TMP_B]] +// OGCG: %[[MUL_AI_BI:.*]] = mul i32 %[[A_IMAG]], 0 +// OGCG: %[[ADD_ARBR_AIBI:.*]] = add i32 %[[MUL_AR_BR]], %[[MUL_AI_BI]] +// OGCG: %[[MUL_BR_BR:.*]] = mul i32 %[[TMP_B]], %[[TMP_B]] +// OGCG: %[[ADD_BRBR_BIBI:.*]] = add i32 %[[MUL_BR_BR]], 0 +// OGCG: %[[MUL_AI_BR:.*]] = mul i32 %[[A_IMAG]], %[[TMP_B]] +// OGCG: %[[MUL_AR_BI:.*]] = mul i32 %[[A_REAL]], 0 +// OGCG: %[[SUB_AIBR_ARBI:.*]] = sub i32 %[[MUL_AI_BR]], %[[MUL_AR_BI]] +// OGCG: %[[RESULT_REAL:.*]] = sdiv i32 %[[ADD_ARBR_AIBI]], %[[ADD_BRBR_BIBI]] +// OGCG: %[[RESULT_IMAG:.*]] = sdiv i32 %[[SUB_AIBR_ARBI]], %[[ADD_BRBR_BIBI]] +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { i32, i32 }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { i32, i32 }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: store i32 %[[RESULT_REAL]], ptr %[[A_REAL_PTR]], align 4 +// OGCG: store i32 %[[RESULT_IMAG]], ptr %[[A_IMAG_PTR]], align 4 + +void foo13() { + _Float16 _Complex a; + _Float16 _Complex b; + b /= (a / b); +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.complex<!cir.f16>, !cir.ptr<!cir.complex<!cir.f16>>, ["a"] +// CIR: %[[B_ADDR:.*]] = cir.alloca !cir.complex<!cir.f16>, !cir.ptr<!cir.complex<!cir.f16>>, ["b"] +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.complex<!cir.f16>>, !cir.complex<!cir.f16> +// CIR: %[[A_REAL:.*]] = cir.complex.real %[[TMP_A]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[A_IMAG:.*]] = cir.complex.imag %[[TMP_A]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[A_REAL_F32:.*]] = cir.cast(floating, %[[A_REAL]] : !cir.f16), !cir.float +// CIR: %[[A_IMAG_F32:.*]] = cir.cast(floating, %[[A_IMAG]] : !cir.f16), !cir.float +// CIR: %[[A_COMPLEX_F32:.*]] = cir.complex.create %[[A_REAL_F32]], %[[A_IMAG_F32]] : !cir.float -> !cir.complex<!cir.float> +// CIR: %[[TMP_B:.*]] = cir.load{{.*}} %[[B_ADDR]] : !cir.ptr<!cir.complex<!cir.f16>>, !cir.complex<!cir.f16> +// CIR: %[[B_REAL:.*]] = cir.complex.real %[[TMP_B]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[B_IMAG:.*]] = cir.complex.imag %[[TMP_B]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[B_REAL_F32:.*]] = cir.cast(floating, %[[B_REAL]] : !cir.f16), !cir.float +// CIR: %[[B_IMAG_F32:.*]] = cir.cast(floating, %[[B_IMAG]] : !cir.f16), !cir.float +// CIR: %[[B_COMPLEX_F32:.*]] = cir.complex.create %[[B_REAL_F32]], %[[B_IMAG_F32]] : !cir.float -> !cir.complex<!cir.float> +// CIR: %[[A_REAL_F32:.*]] = cir.complex.real %[[A_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[A_IMAG_F32:.*]] = cir.complex.imag %[[A_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[B_REAL_F32:.*]] = cir.complex.real %[[B_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[B_IMAG_F32:.*]] = cir.complex.imag %[[B_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[DIV_A_B:.*]] = cir.call @__divsc3(%[[A_REAL_F32]], %[[A_IMAG_F32]], %[[B_REAL_F32]], %[[B_IMAG_F32]]) : (!cir.float, !cir.float, !cir.float, !cir.float) -> !cir.complex<!cir.float> +// CIR: %[[TMP_B:.*]] = cir.load{{.*}} %[[B_ADDR]] : !cir.ptr<!cir.complex<!cir.f16>>, !cir.complex<!cir.f16> +// CIR: %[[B_REAL:.*]] = cir.complex.real %[[TMP_B]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[B_IMAG:.*]] = cir.complex.imag %[[TMP_B]] : !cir.complex<!cir.f16> -> !cir.f16 +// CIR: %[[B_REAL_F32:.*]] = cir.cast(floating, %[[B_REAL]] : !cir.f16), !cir.float +// CIR: %[[B_IMAG_F32:.*]] = cir.cast(floating, %[[B_IMAG]] : !cir.f16), !cir.float +// CIR: %[[B_COMPLEX_F32:.*]] = cir.complex.create %[[B_REAL_F32]], %[[B_IMAG_F32]] : !cir.float -> !cir.complex<!cir.float> +// CIR: %[[B_REAL_F32:.*]] = cir.complex.real %[[B_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[B_IMAG_F32:.*]] = cir.complex.imag %[[B_COMPLEX_F32]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[DIV_AB_REAL:.*]] = cir.complex.real %[[DIV_A_B]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[DIV_AB_IMAG:.*]] = cir.complex.imag %[[DIV_A_B]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[RESULT:.*]] = cir.call @__divsc3(%[[B_REAL_F32]], %[[B_IMAG_F32]], %[[DIV_AB_REAL]], %[[DIV_AB_IMAG]]) : (!cir.float, !cir.float, !cir.float, !cir.float) -> !cir.complex<!cir.float> +// CIR: %[[RESULT_REAL_F32:.*]] = cir.complex.real %[[RESULT]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[RESULT_IMAG_F32:.*]] = cir.complex.imag %[[RESULT]] : !cir.complex<!cir.float> -> !cir.float +// CIR: %[[RESULT_REAL_F16:.*]] = cir.cast(floating, %[[RESULT_REAL_F32]] : !cir.float), !cir.f16 +// CIR: %[[RESULT_IMAG_F16:.*]] = cir.cast(floating, %[[RESULT_IMAG_F32]] : !cir.float), !cir.f16 +// CIR: %[[RESULT_COMPLEX_F16:.*]] = cir.complex.create %[[RESULT_REAL_F16]], %[[RESULT_IMAG_F16]] : !cir.f16 -> !cir.complex<!cir.f16> +// CIR: cir.store{{.*}} %[[RESULT_COMPLEX_F16]], %[[B_ADDR]] : !cir.complex<!cir.f16>, !cir.ptr<!cir.complex<!cir.f16>> + +// LLVM: %[[A_ADDR:.*]] = alloca { half, half }, i64 1, align 2 +// LLVM: %[[B_ADDR:.*]] = alloca { half, half }, i64 1, align 2 +// LLVM: %[[TMP_A:.*]] = load { half, half }, ptr %[[A_ADDR]], align 2 +// LLVM: %[[A_REAL:.*]] = extractvalue { half, half } %[[TMP_A]], 0 +// LLVM: %[[A_IMAG:.*]] = extractvalue { half, half } %[[TMP_A]], 1 +// LLVM: %[[A_REAL_F32:.*]] = fpext half %[[A_REAL]] to float +// LLVM: %[[A_IMAG_F32:.*]] = fpext half %[[A_IMAG]] to float +// LLVM: %[[TMP_A_COMPLEX_F32:.*]] = insertvalue { float, float } {{.*}}, float %[[A_REAL_F32]], 0 +// LLVM: %[[A_COMPLEX_F32:.*]] = insertvalue { float, float } %[[TMP_A_COMPLEX_F32]], float %[[A_IMAG_F32]], 1 +// LLVM: %[[TMP_B:.*]] = load { half, half }, ptr %[[B_ADDR]], align 2 +// LLVM: %[[B_REAL:.*]] = extractvalue { half, half } %[[TMP_B]], 0 +// LLVM: %[[B_IMAG:.*]] = extractvalue { half, half } %[[TMP_B]], 1 +// LLVM: %[[B_REAL_F32:.*]] = fpext half %[[B_REAL]] to float +// LLVM: %[[B_IMAG_F32:.*]] = fpext half %[[B_IMAG]] to float +// LLVM: %[[TMP_B_COMPLEX_F32:.*]] = insertvalue { float, float } {{.*}}, float %[[B_REAL_F32]], 0 +// LLVM: %[[B_COMPLEX_F32:.*]] = insertvalue { float, float } %[[TMP_B_COMPLEX_F32]], float %[[B_IMAG_F32]], 1 +// LLVM: %[[DIV_A_B:.*]] = call { float, float } @__divsc3(float %[[A_REAL_F32]], float %[[A_IMAG_F32]], float %[[B_REAL_F32]], float %[[B_IMAG_F32]]) +// LLVM: %[[TMP_B:.*]] = load { half, half }, ptr %[[B_ADDR]], align 2 +// LLVM: %[[B_REAL:.*]] = extractvalue { half, half } %[[TMP_B]], 0 +// LLVM: %[[B_IMAG:.*]] = extractvalue { half, half } %[[TMP_B]], 1 +// LLVM: %[[B_REAL_F32:.*]] = fpext half %[[B_REAL]] to float +// LLVM: %[[B_IMAG_F32:.*]] = fpext half %[[B_IMAG]] to float +// LLVM: %[[TMP_B_COMPLEX_F32:.*]] = insertvalue { float, float } {{.*}}, float %[[B_REAL_F32]], 0 +// LLVM: %[[B_COMPLEX_F32:.*]] = insertvalue { float, float } %[[TMP_B_COMPLEX_F32]], float %[[B_IMAG_F32]], 1 +// LLVM: %[[DIV_AB_REAL:.*]] = extractvalue { float, float } %[[DIV_A_B]], 0 +// LLVM: %[[DIV_AB_IMAG:.*]] = extractvalue { float, float } %[[DIV_A_B]], 1 +// LLVM: %[[RESULT:.*]] = call { float, float } @__divsc3(float %[[B_REAL_F32]], float %[[B_IMAG_F32]], float %[[DIV_AB_REAL]], float %[[DIV_AB_IMAG]]) +// LLVM: %[[RESULT_REAL_F32:.*]] = extractvalue { float, float } %[[RESULT]], 0 +// LLVM: %[[RESULT_IMAG_F32:.*]] = extractvalue { float, float } %[[RESULT]], 1 +// LLVM: %[[RESULT_REAL_F16:.*]] = fptrunc float %[[RESULT_REAL_F32]] to half +// LLVM: %[[RESULT_IMAG_F16:.*]] = fptrunc float %[[RESULT_IMAG_F32]] to half +// LLVM: %[[TMP_RESULT_F16:.*]] = insertvalue { half, half } {{.*}}, half %[[RESULT_REAL_F16]], 0 +// LLVM: %[[RESULT_COMPLEX_F16:.*]] = insertvalue { half, half } %[[TMP_RESULT_F16]], half %[[RESULT_IMAG_F16]], 1 +// LLVM: store { half, half } %[[RESULT_COMPLEX_F16]], ptr %[[B_ADDR]], align 2 + +// OGCG: %[[A_ADDR:.*]] = alloca { half, half }, align 2 +// OGCG: %[[B_ADDR:.*]] = alloca { half, half }, align 2 +// OGCG: %[[DIV_AB_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[DIV_B_AB_ADDR:.*]] = alloca { float, float }, align 4 +// OGCG: %[[A_REAL_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[A_ADDR]], i32 0, i32 0 +// OGCG: %[[A_REAL:.*]] = load half, ptr %[[A_REAL_PTR]], align 2 +// OGCG: %[[A_IMAG_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[A_ADDR]], i32 0, i32 1 +// OGCG: %[[A_IMAG:.*]] = load half, ptr %[[A_IMAG_PTR]], align 2 +// OGCG: %[[A_REAL_F32:.*]] = fpext half %[[A_REAL]] to float +// OGCG: %[[A_IMAG_F32:.*]] = fpext half %[[A_IMAG]] to float +// OGCG: %[[B_REAL_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 0 +// OGCG: %[[B_REAL:.*]] = load half, ptr %[[B_REAL_PTR]], align 2 +// OGCG: %[[B_IMAG_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 1 +// OGCG: %[[B_IMAG:.*]] = load half, ptr %[[B_IMAG_PTR]], align 2 +// OGCG: %[[B_REAL_F32:.*]] = fpext half %[[B_REAL]] to float +// OGCG: %[[B_IMAG_F32:.*]] = fpext half %[[B_IMAG]] to float +// OGCG: %[[DIV_A_B:.*]] = call{{.*}} <2 x float> @__divsc3(float noundef %[[A_REAL_F32]], float noundef %[[A_IMAG_F32]], float noundef %[[B_REAL_F32]], float noundef %[[B_IMAG_F32]]) +// OGCG: store <2 x float> %[[DIV_A_B]], ptr %[[DIV_AB_ADDR]], align 4 +// OGCG: %[[DIV_AB_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[DIV_AB_ADDR]], i32 0, i32 0 +// OGCG: %[[DIV_AB_REAL:.*]] = load float, ptr %[[DIV_AB_REAL_PTR]], align 4 +// OGCG: %[[DIV_AB_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[DIV_AB_ADDR]], i32 0, i32 1 +// OGCG: %[[DIV_AB_IMAG:.*]] = load float, ptr %[[DIV_AB_IMAG_PTR]], align 4 +// OGCG: %[[B_REAL_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 0 +// OGCG: %[[B_REAL:.*]] = load half, ptr %[[B_REAL_PTR]], align 2 +// OGCG: %[[B_IMAG_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 1 +// OGCG: %[[B_IMAG:.*]] = load half, ptr %[[B_IMAG_PTR]], align 2 +// OGCG: %[[B_REAL_F32:.*]] = fpext half %[[B_REAL]] to float +// OGCG: %[[B_IMAG_F32:.*]] = fpext half %[[B_IMAG]] to float +// OGCG: %[[RESULT:.*]] = call{{.*}} <2 x float> @__divsc3(float noundef %[[B_REAL_F32]], float noundef %[[B_IMAG_F32]], float noundef %[[DIV_AB_REAL]], float noundef %[[DIV_AB_IMAG]]) +// OGCG: store <2 x float> %[[RESULT]], ptr %[[DIV_B_AB_ADDR]], align 4 +// OGCG: %[[RESULT_REAL_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[DIV_B_AB_ADDR]], i32 0, i32 0 +// OGCG: %[[RESULT_REAL:.*]] = load float, ptr %[[RESULT_REAL_PTR]], align 4 +// OGCG: %[[RESULT_IMAG_PTR:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[DIV_B_AB_ADDR]], i32 0, i32 1 +// OGCG: %[[RESULT_IMAG:.*]] = load float, ptr %[[RESULT_IMAG_PTR]], align 4 +// OGCG: %[[RESULT_REAL_F16:.*]] = fptrunc float %[[RESULT_REAL]] to half +// OGCG: %[[RESULT_IMAG_F16:.*]] = fptrunc float %[[RESULT_IMAG]] to half +// OGCG: %[[B_REAL_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 0 +// OGCG: %[[B_IMAG_PTR:.*]] = getelementptr inbounds nuw { half, half }, ptr %[[B_ADDR]], i32 0, i32 1 +// OGCG: store half %[[RESULT_REAL_F16]], ptr %[[B_REAL_PTR]], align 2 +// OGCG: store half %[[RESULT_IMAG_F16]], ptr %[[B_IMAG_PTR]], align 2 + #ifndef __cplusplus void foo9() { float _Complex a; |
