Skip to content

[C2y] Claim conformance to WG14 N3460 #131196

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Mar 14, 2025

Conversation

AaronBallman
Copy link
Collaborator

This moves some Annex G requirements for complex numbers into the main
body of the standard.

This moves some Annex G requirements for complex numbers into the main
body of the standard.
@AaronBallman AaronBallman added clang:frontend Language frontend issues, e.g. anything involving "Sema" floating-point Floating-point math c2y labels Mar 13, 2025
@llvmbot llvmbot added the clang Clang issues not falling into any other category label Mar 13, 2025
@llvmbot
Copy link
Member

llvmbot commented Mar 13, 2025

@llvm/pr-subscribers-clang

Author: Aaron Ballman (AaronBallman)

Changes

This moves some Annex G requirements for complex numbers into the main
body of the standard.


Full diff: https://github.com/llvm/llvm-project/pull/131196.diff

3 Files Affected:

  • (added) clang/test/C/C2y/n3460.c (+51)
  • (added) clang/test/C/C2y/n3460_1.c (+80)
  • (modified) clang/www/c_status.html (+1-1)
diff --git a/clang/test/C/C2y/n3460.c b/clang/test/C/C2y/n3460.c
new file mode 100644
index 0000000000000..e21c60b418721
--- /dev/null
+++ b/clang/test/C/C2y/n3460.c
@@ -0,0 +1,51 @@
+// RUN: %clang_cc1 -verify -std=c2y -Wall %s
+
+/* WG14 N3460: Clang 12
+ * Complex operators
+ *
+ * This moves some Annex G requirements into the main body of the standard.
+ */
+
+// CMPLX(0.0, inf) * 2.0, the result should be CMPLX(0.0, inf), not CMPLX(nan, inf)
+static_assert(__builtin_complex(0.0, __builtin_inf()) * 2.0 ==
+              __builtin_complex(0.0, __builtin_inf()));
+
+// CMPLX(0.0, 1.0) * -0.0 is CMPLX(-0.0, -0.0), not CMPLX(-0.0, +0.0)
+static_assert(__builtin_complex(0.0, 1.0) * -0.0 ==
+              __builtin_complex(-0.0, -0.0));
+
+// Testing for -0.0 is a pain because -0.0 == +0.0, so forcefully generate a
+// diagnostic and check the note.
+static_assert(__builtin_complex(0.0, 1.0) * -0.0 == 1); /* expected-error {{static assertion failed due to requirement '__builtin_complex(0., 1.) * -0. == 1'}} \
+                                                           expected-note {{expression evaluates to '(-0 + -0i) == 1'}}
+                                                         */
+
+// CMPLX(0.0, inf) / 2.0, the result should be CMPLX(0.0, inf),
+// not CMPLX(nan, inf)
+static_assert(__builtin_complex(0.0, __builtin_inf()) / 2.0 ==
+              __builtin_complex(0.0, __builtin_inf()));
+
+// CMPLX(2.0, 3.0) * 2.0, the result should be CMPLX(4.0, 6.0)
+static_assert(__builtin_complex(2.0, 3.0) * 2.0 ==
+              __builtin_complex(4.0, 6.0));
+
+// CMPLX(2.0, 4.0) / 2.0, the result should be CMPLX(1.0, 2.0)
+static_assert(__builtin_complex(2.0, 4.0) / 2.0 ==
+              __builtin_complex(1.0, 2.0));
+
+// CMPLX(2.0, 3.0) * CMPLX(4.0, 5.0), the result should be
+// CMPLX(8.0 - 15.0, 12.0 + 10.0)
+static_assert(__builtin_complex(2.0, 3.0) * __builtin_complex(4.0, 5.0) ==
+              __builtin_complex(-7.0, 22.0));
+
+// CMPLX(2.0, 3.0) / CMPLX(4.0, 5.0), the result should be
+// CMPLX((8.0 + 15.0)/(4.0^2 + 5.0^2), (12.0 - 10.0)/(4.0^2 + 5.0^2))
+static_assert(__builtin_complex(2.0, 3.0) / __builtin_complex(4.0, 5.0) ==
+              __builtin_complex(23.0 / 41.0, 2.0 / 41.0));
+
+
+// 2.0 / CMPLX(2.0, 4.0), the result should be
+// CMPLX(4.0 /(2.0^2 + 4.0^2), -8.0/(2.0^2 + 4.0^2))
+static_assert(2.0 / __builtin_complex(2.0, 4.0) ==
+              __builtin_complex(4.0 / 20.0, -8.0 / 20.0));
+
diff --git a/clang/test/C/C2y/n3460_1.c b/clang/test/C/C2y/n3460_1.c
new file mode 100644
index 0000000000000..2d50b3c23dae8
--- /dev/null
+++ b/clang/test/C/C2y/n3460_1.c
@@ -0,0 +1,80 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// RUN: %clang_cc1 -std=c2y -O0 -triple x86_64-unknown-unknown %s -emit-llvm -o - | FileCheck %s
+// This tests the codegen for the same test cases as in n3460.c.
+
+// CHECK-LABEL: define dso_local void @test(
+// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT:  [[ENTRY:.*]]:
+// CHECK-NEXT:    [[A:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[B:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[C:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[D:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[E:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[F:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[G:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[H:%.*]] = alloca { double, double }, align 8
+// CHECK-NEXT:    [[A_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[A]], i32 0, i32 0
+// CHECK-NEXT:    [[A_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[A]], i32 0, i32 1
+// CHECK-NEXT:    store double 0.000000e+00, ptr [[A_REALP]], align 8
+// CHECK-NEXT:    store double 0x7FF0000000000000, ptr [[A_IMAGP]], align 8
+// CHECK-NEXT:    [[B_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[B]], i32 0, i32 0
+// CHECK-NEXT:    [[B_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[B]], i32 0, i32 1
+// CHECK-NEXT:    store double -0.000000e+00, ptr [[B_REALP]], align 8
+// CHECK-NEXT:    store double -0.000000e+00, ptr [[B_IMAGP]], align 8
+// CHECK-NEXT:    [[C_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[C]], i32 0, i32 0
+// CHECK-NEXT:    [[C_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[C]], i32 0, i32 1
+// CHECK-NEXT:    store double 0.000000e+00, ptr [[C_REALP]], align 8
+// CHECK-NEXT:    store double 0x7FF0000000000000, ptr [[C_IMAGP]], align 8
+// CHECK-NEXT:    [[D_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[D]], i32 0, i32 0
+// CHECK-NEXT:    [[D_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[D]], i32 0, i32 1
+// CHECK-NEXT:    store double 4.000000e+00, ptr [[D_REALP]], align 8
+// CHECK-NEXT:    store double 6.000000e+00, ptr [[D_IMAGP]], align 8
+// CHECK-NEXT:    [[E_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[E]], i32 0, i32 0
+// CHECK-NEXT:    [[E_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[E]], i32 0, i32 1
+// CHECK-NEXT:    store double 1.000000e+00, ptr [[E_REALP]], align 8
+// CHECK-NEXT:    store double 2.000000e+00, ptr [[E_IMAGP]], align 8
+// CHECK-NEXT:    br i1 false, label %[[COMPLEX_MUL_IMAG_NAN:.*]], label %[[COMPLEX_MUL_CONT:.*]], !prof [[PROF2:![0-9]+]]
+// CHECK:       [[COMPLEX_MUL_IMAG_NAN]]:
+// CHECK-NEXT:    br i1 false, label %[[COMPLEX_MUL_LIBCALL:.*]], label %[[COMPLEX_MUL_CONT]], !prof [[PROF2]]
+// CHECK:       [[COMPLEX_MUL_LIBCALL]]:
+// CHECK-NEXT:    [[CALL:%.*]] = call { double, double } @__muldc3(double noundef 2.000000e+00, double noundef 3.000000e+00, double noundef 4.000000e+00, double noundef 5.000000e+00) #[[ATTR1:[0-9]+]]
+// CHECK-NEXT:    [[TMP0:%.*]] = extractvalue { double, double } [[CALL]], 0
+// CHECK-NEXT:    [[TMP1:%.*]] = extractvalue { double, double } [[CALL]], 1
+// CHECK-NEXT:    br label %[[COMPLEX_MUL_CONT]]
+// CHECK:       [[COMPLEX_MUL_CONT]]:
+// CHECK-NEXT:    [[REAL_MUL_PHI:%.*]] = phi double [ -7.000000e+00, %[[ENTRY]] ], [ -7.000000e+00, %[[COMPLEX_MUL_IMAG_NAN]] ], [ [[TMP0]], %[[COMPLEX_MUL_LIBCALL]] ]
+// CHECK-NEXT:    [[IMAG_MUL_PHI:%.*]] = phi double [ 2.200000e+01, %[[ENTRY]] ], [ 2.200000e+01, %[[COMPLEX_MUL_IMAG_NAN]] ], [ [[TMP1]], %[[COMPLEX_MUL_LIBCALL]] ]
+// CHECK-NEXT:    [[F_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[F]], i32 0, i32 0
+// CHECK-NEXT:    [[F_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[F]], i32 0, i32 1
+// CHECK-NEXT:    store double [[REAL_MUL_PHI]], ptr [[F_REALP]], align 8
+// CHECK-NEXT:    store double [[IMAG_MUL_PHI]], ptr [[F_IMAGP]], align 8
+// CHECK-NEXT:    [[CALL1:%.*]] = call { double, double } @__divdc3(double noundef 2.000000e+00, double noundef 3.000000e+00, double noundef 4.000000e+00, double noundef 5.000000e+00) #[[ATTR1]]
+// CHECK-NEXT:    [[TMP2:%.*]] = extractvalue { double, double } [[CALL1]], 0
+// CHECK-NEXT:    [[TMP3:%.*]] = extractvalue { double, double } [[CALL1]], 1
+// CHECK-NEXT:    [[G_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[G]], i32 0, i32 0
+// CHECK-NEXT:    [[G_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[G]], i32 0, i32 1
+// CHECK-NEXT:    store double [[TMP2]], ptr [[G_REALP]], align 8
+// CHECK-NEXT:    store double [[TMP3]], ptr [[G_IMAGP]], align 8
+// CHECK-NEXT:    [[CALL2:%.*]] = call { double, double } @__divdc3(double noundef 2.000000e+00, double noundef 0.000000e+00, double noundef 2.000000e+00, double noundef 4.000000e+00) #[[ATTR1]]
+// CHECK-NEXT:    [[TMP4:%.*]] = extractvalue { double, double } [[CALL2]], 0
+// CHECK-NEXT:    [[TMP5:%.*]] = extractvalue { double, double } [[CALL2]], 1
+// CHECK-NEXT:    [[H_REALP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[H]], i32 0, i32 0
+// CHECK-NEXT:    [[H_IMAGP:%.*]] = getelementptr inbounds nuw { double, double }, ptr [[H]], i32 0, i32 1
+// CHECK-NEXT:    store double [[TMP4]], ptr [[H_REALP]], align 8
+// CHECK-NEXT:    store double [[TMP5]], ptr [[H_IMAGP]], align 8
+// CHECK-NEXT:    ret void
+//
+void test() {
+  _Complex double a  = __builtin_complex(0.0, __builtin_inf()) * 2.0;
+  _Complex double b = __builtin_complex(0.0, 1.0) * -0.0;
+  _Complex double c = __builtin_complex(0.0, __builtin_inf()) / 2.0;
+  _Complex double d = __builtin_complex(2.0, 3.0) * 2.0;
+  _Complex double e = __builtin_complex(2.0, 4.0) / 2.0;
+  _Complex double f = __builtin_complex(2.0, 3.0) * __builtin_complex(4.0, 5.0);
+  _Complex double g = __builtin_complex(2.0, 3.0) / __builtin_complex(4.0, 5.0);
+  _Complex double h = 2.0 / __builtin_complex(2.0, 4.0);
+}
+
+//.
+// CHECK: [[PROF2]] = !{!"branch_weights", i32 1, i32 1048575}
+//.
diff --git a/clang/www/c_status.html b/clang/www/c_status.html
index f3f9ffadae7ec..d68e8d6441ed2 100644
--- a/clang/www/c_status.html
+++ b/clang/www/c_status.html
@@ -318,7 +318,7 @@ <h2 id="c2y">C2y implementation status</h2>
     <tr>
       <td>Complex operators</td>
       <td><a href="https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3460.pdf">N3460</a></td>
-      <td class="unknown" align="center">Unknown</td>
+      <td class="full" align="center">Clang 12</td>
 	</tr>
 </table>
 </details>

@AaronBallman AaronBallman merged commit cbbcc3d into llvm:main Mar 14, 2025
16 checks passed
@AaronBallman AaronBallman deleted the aballman-wg14-n3460 branch March 14, 2025 16:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
c2y clang:frontend Language frontend issues, e.g. anything involving "Sema" clang Clang issues not falling into any other category floating-point Floating-point math
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants