Skip to content

[libc][math][c23] Add {totalorder,totalordermag}f16 C23 math functions #95014

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.

Sign up for GitHub

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

Jump to bottom
Merged
merged 4 commits into from
Jun 11, 2024
Merged

[libc][math][c23] Add {totalorder,totalordermag}f16 C23 math functions #95014

merged 4 commits into from
Jun 11, 2024

Conversation

overmighty
Copy link
Member

@overmighty overmighty commented Jun 10, 2024 •
edited
Loading

Part of #93566.

@overmighty
Copy link
Member Author

cc @lntue

@llvmbot llvmbot added the libc label Jun 10, 2024
@llvmbot
Copy link
Member

llvmbot commented Jun 10, 2024

@llvm/pr-subscribers-libc

Author: OverMighty (overmighty)

Changes

Part of #93566.

Patch is 27.32 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/95014.diff

19 Files Affected:

  • (modified) libc/config/linux/aarch64/entrypoints.txt (+2)
  • (modified) libc/config/linux/x86_64/entrypoints.txt (+2)
  • (modified) libc/docs/c23.rst (+2-2)
  • (modified) libc/docs/math/index.rst (+4)
  • (modified) libc/spec/stdc.td (+4)
  • (modified) libc/src/__support/FPUtil/BasicOperations.h (+51)
  • (modified) libc/src/__support/FPUtil/FPBits.h (+4)
  • (modified) libc/src/math/CMakeLists.txt (+4)
  • (modified) libc/src/math/generic/CMakeLists.txt (+24)
  • (added) libc/src/math/generic/totalorderf16.cpp (+19)
  • (added) libc/src/math/generic/totalordermagf16.cpp (+20)
  • (added) libc/src/math/totalorderf16.h (+20)
  • (added) libc/src/math/totalordermagf16.h (+20)
  • (modified) libc/test/UnitTest/FPMatcher.h (+4-2)
  • (modified) libc/test/src/math/smoke/CMakeLists.txt (+24)
  • (added) libc/test/src/math/smoke/TotalOrderMagTest.h (+143)
  • (added) libc/test/src/math/smoke/TotalOrderTest.h (+139)
  • (added) libc/test/src/math/smoke/totalorderf16_test.cpp (+13)
  • (added) libc/test/src/math/smoke/totalordermagf16_test.cpp (+13)
diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt
index 8863749e12c6d..dba4835afb745 100644
--- a/libc/config/linux/aarch64/entrypoints.txt
+++ b/libc/config/linux/aarch64/entrypoints.txt
@@ -539,6 +539,8 @@ if(LIBC_TYPES_HAS_FLOAT16)
     libc.src.math.rintf16
     libc.src.math.roundf16
     libc.src.math.roundevenf16
+    libc.src.math.totalorderf16
+    libc.src.math.totalordermagf16
     libc.src.math.truncf16
     libc.src.math.ufromfpf16
     libc.src.math.ufromfpxf16
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index 31ad0bc412836..90505738d624d 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -569,6 +569,8 @@ if(LIBC_TYPES_HAS_FLOAT16)
     libc.src.math.rintf16
     libc.src.math.roundf16
     libc.src.math.roundevenf16
+    libc.src.math.totalorderf16
+    libc.src.math.totalordermagf16
     libc.src.math.truncf16
     libc.src.math.ufromfpf16
     libc.src.math.ufromfpxf16
diff --git a/libc/docs/c23.rst b/libc/docs/c23.rst
index fec9b24bbd581..4134befd1ed35 100644
--- a/libc/docs/c23.rst
+++ b/libc/docs/c23.rst
@@ -42,8 +42,8 @@ Additions:
   * rsqrt*
   * __STDC_IEC_60559_DFP__ functions (_Decimal32, _Decimal64, _Decimal128)
   * compoundn*
-  * totalorder*
-  * totalordermag*
+  * totalorder* |check|
+  * totalordermag* |check|
   * getpayload*
   * setpayload*
   * iscannonical
diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst
index 3e122fb8bc26e..953bdd8b642ad 100644
--- a/libc/docs/math/index.rst
+++ b/libc/docs/math/index.rst
@@ -210,6 +210,10 @@ Basic Operations
 +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
 | scalbn           | |check|          | |check|         | |check|                |                      | |check|                | 7.12.6.19              | F.10.3.19                  |
 +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
+| totalorder       |                  |                 |                        | |check|              |                        | F.10.12.1              | N/A                        |
++------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
+| totalordermag    |                  |                 |                        | |check|              |                        | F.10.12.2              | N/A                        |
++------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
 | trunc            | |check|          | |check|         | |check|                | |check|              | |check|                | 7.12.9.9               | F.10.6.9                   |
 +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
 | ufromfp          | |check|          | |check|         | |check|                | |check|              | |check|                | 7.12.9.10              | F.10.6.10                  |
diff --git a/libc/spec/stdc.td b/libc/spec/stdc.td
index b5b6dbc481bd7..b1b02a8e1ed31 100644
--- a/libc/spec/stdc.td
+++ b/libc/spec/stdc.td
@@ -708,6 +708,10 @@ def StdC : StandardSpec<"stdc"> {
           FunctionSpec<"canonicalizel", RetValSpec<IntType>, [ArgSpec<LongDoubleType>, ArgSpec<LongDoubleType>]>,
           GuardedFunctionSpec<"canonicalizef16", RetValSpec<IntType>, [ArgSpec<Float16Type>, ArgSpec<Float16Type>], "LIBC_TYPES_HAS_FLOAT16">,
           GuardedFunctionSpec<"canonicalizef128", RetValSpec<IntType>, [ArgSpec<Float128Type>, ArgSpec<Float128Type>], "LIBC_TYPES_HAS_FLOAT128">,
+
+          GuardedFunctionSpec<"totalorderf16", RetValSpec<IntType>, [ArgSpec<Float16Ptr>, ArgSpec<Float16Ptr>], "LIBC_TYPES_HAS_FLOAT16">,
+
+          GuardedFunctionSpec<"totalordermagf16", RetValSpec<IntType>, [ArgSpec<Float16Ptr>, ArgSpec<Float16Ptr>], "LIBC_TYPES_HAS_FLOAT16">,
       ]
   >;
 
diff --git a/libc/src/__support/FPUtil/BasicOperations.h b/libc/src/__support/FPUtil/BasicOperations.h
index e5ac101fedc0e..168b3d308f0ee 100644
--- a/libc/src/__support/FPUtil/BasicOperations.h
+++ b/libc/src/__support/FPUtil/BasicOperations.h
@@ -240,6 +240,57 @@ LIBC_INLINE int canonicalize(T &cx, const T &x) {
   return 0;
 }
 
+template <typename T>
+LIBC_INLINE cpp::enable_if_t<cpp::is_floating_point_v<T>, bool>
+totalorder(const T *x, const T *y) {
+  using FPBits = FPBits<T>;
+  FPBits xbits(*x);
+  FPBits ybits(*y);
+
+  if (LIBC_UNLIKELY(xbits.is_zero() && ybits.is_zero() || xbits.is_nan() ||
+                    ybits.is_nan())) {
+    // totalOrder(−0, +0) is true. totalOrder(+0, -0) is false.
+    if (xbits.is_zero() && ybits.is_zero()) {
+      Sign xsign = xbits.sign();
+      Sign ysign = ybits.sign();
+      return (xsign == Sign::NEG && ysign == Sign::POS) || xsign == ysign;
+    }
+
+    // totalOrder(−NaN, y) is true. totalOrder(+NaN, y) is false.
+    if (!ybits.is_nan())
+      return xbits.is_neg();
+    // totalOrder(x, +NaN) is true. totalOrder(x, -NaN) is false.
+    if (!xbits.is_nan())
+      return ybits.is_pos();
+
+    // Negative sign orders below positive sign.
+    if (xbits.is_neg() && ybits.is_pos())
+      return true;
+    if (xbits.is_pos() && ybits.is_neg())
+      return false;
+
+    // Signaling orders below quiet for +NaN, reverse for −NaN.
+    if (xbits.is_quiet_nan() && ybits.is_signaling_nan())
+      return xbits.is_neg();
+    if (xbits.is_signaling_nan() && ybits.is_quiet_nan())
+      return ybits.is_pos();
+
+    // Otherwise, the order of NaNs is implementation-defined (IEEE 754-2019).
+    // We order by payload.
+    return xbits.get_nan_payload() <= ybits.get_nan_payload();
+  }
+
+  return xbits.get_val() <= ybits.get_val();
+}
+
+template <typename T>
+LIBC_INLINE cpp::enable_if_t<cpp::is_floating_point_v<T>, bool>
+totalordermag(const T *x, const T *y) {
+  T abs_x = abs(*x);
+  T abs_y = abs(*y);
+  return totalorder(&abs_x, &abs_y);
+}
+
 } // namespace fputil
 } // namespace LIBC_NAMESPACE
 
diff --git a/libc/src/__support/FPUtil/FPBits.h b/libc/src/__support/FPUtil/FPBits.h
index 559ecde767c30..f424e11ec9fd1 100644
--- a/libc/src/__support/FPUtil/FPBits.h
+++ b/libc/src/__support/FPUtil/FPBits.h
@@ -708,6 +708,10 @@ struct FPRepImpl : public FPRepSem<fp_type, RetT> {
   LIBC_INLINE constexpr void set_significand(StorageType sigVal) {
     bits = UP::merge(bits, sigVal, SIG_MASK);
   }
+
+  LIBC_INLINE constexpr StorageType get_nan_payload() const {
+    return bits & (FRACTION_MASK >> 1);
+  }
   // Unsafe function to create a floating point representation.
   // It simply packs the sign, biased exponent and mantissa values without
   // checking bound nor normalization.
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index f8582d8d42683..3c1a5c86f6b48 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -367,6 +367,10 @@ add_math_entrypoint_object(tanhf)
 add_math_entrypoint_object(tgamma)
 add_math_entrypoint_object(tgammaf)
 
+add_math_entrypoint_object(totalorderf16)
+
+add_math_entrypoint_object(totalordermagf16)
+
 add_math_entrypoint_object(trunc)
 add_math_entrypoint_object(truncf)
 add_math_entrypoint_object(truncl)
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index caaa0ac23dc7a..a13cadc8acd0c 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -3551,3 +3551,27 @@ add_entrypoint_object(
   COMPILE_OPTIONS
     -O3
 )
+
+add_entrypoint_object(
+  totalorderf16
+  SRCS
+    totalorderf16.cpp
+  HDRS
+    ../totalorderf16.h
+  DEPENDS
+    libc.src.__support.FPUtil.basic_operations
+  COMPILE_OPTIONS
+    -O3
+)
+
+add_entrypoint_object(
+  totalordermagf16
+  SRCS
+    totalordermagf16.cpp
+  HDRS
+    ../totalordermagf16.h
+  DEPENDS
+    libc.src.__support.FPUtil.basic_operations
+  COMPILE_OPTIONS
+    -O3
+)
diff --git a/libc/src/math/generic/totalorderf16.cpp b/libc/src/math/generic/totalorderf16.cpp
new file mode 100644
index 0000000000000..b64f1eeaf3ddf
--- /dev/null
+++ b/libc/src/math/generic/totalorderf16.cpp
@@ -0,0 +1,19 @@
+//===-- Implementation of totalorderf16 function --------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/totalorderf16.h"
+#include "src/__support/FPUtil/BasicOperations.h"
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(int, totalorderf16, (const float16 *x, const float16 *y)) {
+  return static_cast<int>(fputil::totalorder(x, y));
+}
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/generic/totalordermagf16.cpp b/libc/src/math/generic/totalordermagf16.cpp
new file mode 100644
index 0000000000000..868e155bac32b
--- /dev/null
+++ b/libc/src/math/generic/totalordermagf16.cpp
@@ -0,0 +1,20 @@
+//===-- Implementation of totalordermagf16 function -----------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/totalordermagf16.h"
+#include "src/__support/FPUtil/BasicOperations.h"
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(int, totalordermagf16,
+                   (const float16 *x, const float16 *y)) {
+  return static_cast<int>(fputil::totalordermag(x, y));
+}
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/totalorderf16.h b/libc/src/math/totalorderf16.h
new file mode 100644
index 0000000000000..f5390140c4dc2
--- /dev/null
+++ b/libc/src/math/totalorderf16.h
@@ -0,0 +1,20 @@
+//===-- Implementation header for totalorderf16 -----------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_MATH_TOTALORDERF16_H
+#define LLVM_LIBC_SRC_MATH_TOTALORDERF16_H
+
+#include "src/__support/macros/properties/types.h"
+
+namespace LIBC_NAMESPACE {
+
+int totalorderf16(const float16 *x, const float16 *y);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_MATH_TOTALORDERF16_H
diff --git a/libc/src/math/totalordermagf16.h b/libc/src/math/totalordermagf16.h
new file mode 100644
index 0000000000000..8c6621b9783df
--- /dev/null
+++ b/libc/src/math/totalordermagf16.h
@@ -0,0 +1,20 @@
+//===-- Implementation header for totalordermagf16 --------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_MATH_TOTALORDERMAGF16_H
+#define LLVM_LIBC_SRC_MATH_TOTALORDERMAGF16_H
+
+#include "src/__support/macros/properties/types.h"
+
+namespace LIBC_NAMESPACE {
+
+int totalordermagf16(const float16 *x, const float16 *y);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_MATH_TOTALORDERMAGF16_H
diff --git a/libc/test/UnitTest/FPMatcher.h b/libc/test/UnitTest/FPMatcher.h
index 26af5cec02b58..86b823260e197 100644
--- a/libc/test/UnitTest/FPMatcher.h
+++ b/libc/test/UnitTest/FPMatcher.h
@@ -97,8 +97,10 @@ template <typename T> struct FPTest : public Test {
       LIBC_NAMESPACE::cpp::numeric_limits<StorageType>::max();                 \
   const T zero = FPBits::zero(Sign::POS).get_val();                            \
   const T neg_zero = FPBits::zero(Sign::NEG).get_val();                        \
-  const T aNaN = FPBits::quiet_nan().get_val();                                \
-  const T sNaN = FPBits::signaling_nan().get_val();                            \
+  const T aNaN = FPBits::quiet_nan(Sign::POS).get_val();                       \
+  const T neg_aNaN = FPBits::quiet_nan(Sign::NEG).get_val();                   \
+  const T sNaN = FPBits::signaling_nan(Sign::POS).get_val();                   \
+  const T neg_sNaN = FPBits::signaling_nan(Sign::NEG).get_val();               \
   const T inf = FPBits::inf(Sign::POS).get_val();                              \
   const T neg_inf = FPBits::inf(Sign::NEG).get_val();                          \
   const T min_normal = FPBits::min_normal().get_val();                         \
diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt
index 84aa76c0a0881..706e23518ed58 100644
--- a/libc/test/src/math/smoke/CMakeLists.txt
+++ b/libc/test/src/math/smoke/CMakeLists.txt
@@ -3509,3 +3509,27 @@ add_fp_unittest(
     libc.src.math.powf
     libc.src.__support.FPUtil.fp_bits
 )
+
+add_fp_unittest(
+  totalorderf16_test
+  SUITE
+    libc-math-smoke-tests
+  SRCS
+    totalorderf16_test.cpp
+  HDRS
+    TotalOrderTest.h
+  DEPENDS
+    libc.src.math.totalorderf16
+)
+
+add_fp_unittest(
+  totalordermagf16_test
+  SUITE
+    libc-math-smoke-tests
+  SRCS
+    totalordermagf16_test.cpp
+  HDRS
+    TotalOrderMagTest.h
+  DEPENDS
+    libc.src.math.totalordermagf16
+)
diff --git a/libc/test/src/math/smoke/TotalOrderMagTest.h b/libc/test/src/math/smoke/TotalOrderMagTest.h
new file mode 100644
index 0000000000000..013a7fc92f9b1
--- /dev/null
+++ b/libc/test/src/math/smoke/TotalOrderMagTest.h
@@ -0,0 +1,143 @@
+//===-- Utility class to test flavors of totalordermag ----------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LIBC_TEST_SRC_MATH_SMOKE_TOTALORDERMAGTEST_H
+#define LIBC_TEST_SRC_MATH_SMOKE_TOTALORDERMAGTEST_H
+
+#include "test/UnitTest/FEnvSafeTest.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+
+template <typename T>
+class TotalOrderMagTestTemplate : public LIBC_NAMESPACE::testing::FEnvSafeTest {
+
+  DECLARE_SPECIAL_CONSTANTS(T)
+
+public:
+  typedef int (*TotalOrderMagFunc)(const T *, const T *);
+
+  bool funcWrapper(TotalOrderMagFunc func, T x, T y) {
+    return func(&x, &y) != 0;
+  }
+
+  void testXLesserThanY(TotalOrderMagFunc func) {
+    EXPECT_TRUE(funcWrapper(func, neg_inf, inf));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(123.38)));
+
+    EXPECT_FALSE(funcWrapper(func, T(-0.1), T(0.0)));
+    EXPECT_FALSE(funcWrapper(func, T(-123.38), T(0.0)));
+
+    EXPECT_TRUE(funcWrapper(func, T(-0.1), T(0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(-123.38), T(123.38)));
+  }
+
+  void testXGreaterThanY(TotalOrderMagFunc func) {
+    EXPECT_TRUE(funcWrapper(func, inf, neg_inf));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(-0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(-123.38)));
+
+    EXPECT_FALSE(funcWrapper(func, T(0.1), T(0.0)));
+    EXPECT_FALSE(funcWrapper(func, T(123.38), T(0.0)));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.1), T(-0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(123.38), T(-123.38)));
+  }
+
+  void testXEqualToY(TotalOrderMagFunc func) {
+    EXPECT_TRUE(funcWrapper(func, inf, inf));
+    EXPECT_TRUE(funcWrapper(func, neg_inf, neg_inf));
+
+    EXPECT_TRUE(funcWrapper(func, T(-0.0), T(0.0)));
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(-0.0)));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.0), T(0.0)));
+    EXPECT_TRUE(funcWrapper(func, T(-0.0), T(-0.0)));
+    EXPECT_TRUE(funcWrapper(func, T(0.1), T(0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(-0.1), T(-0.1)));
+    EXPECT_TRUE(funcWrapper(func, T(123.38), T(123.38)));
+    EXPECT_TRUE(funcWrapper(func, T(-123.38), T(-123.38)));
+  }
+
+  void testSingleNaN(TotalOrderMagFunc func) {
+    EXPECT_FALSE(funcWrapper(func, neg_aNaN, T(0.0)));
+    EXPECT_FALSE(funcWrapper(func, neg_aNaN, T(0.1)));
+    EXPECT_FALSE(funcWrapper(func, neg_aNaN, T(123.38)));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.0), neg_aNaN));
+    EXPECT_TRUE(funcWrapper(func, T(0.1), neg_aNaN));
+    EXPECT_TRUE(funcWrapper(func, T(123.38), neg_aNaN));
+
+    EXPECT_TRUE(funcWrapper(func, T(0.0), aNaN));
+    EXPECT_TRUE(funcWrapper(func, T(0.1), aNaN));
+    EXPECT_TRUE(funcWrapper(func, T(123.38), aNaN));
+
+    EXPECT_FALSE(funcWrapper(func, aNaN, T(0.0)));
+    EXPECT_FALSE(funcWrapper(func, aNaN, T(0.1)));
+    EXPECT_FALSE(funcWrapper(func, aNaN, T(123.38)));
+  }
+
+  void testNaNSigns(TotalOrderMagFunc func) {
+    EXPECT_TRUE(funcWrapper(func, neg_aNaN, aNaN));
+    EXPECT_FALSE(funcWrapper(func, neg_aNaN, sNaN));
+    EXPECT_TRUE(funcWrapper(func, neg_sNaN, aNaN));
+    EXPECT_TRUE(funcWrapper(func, neg_sNaN, sNaN));
+
+    EXPECT_TRUE(funcWrapper(func, aNaN, neg_aNaN));
+    EXPECT_FALSE(funcWrapper(func, aNaN, neg_sNaN));
+    EXPECT_TRUE(funcWrapper(func, sNaN, neg_aNaN));
+    EXPECT_TRUE(funcWrapper(func, sNaN, neg_sNaN));
+  }
+
+  void testQuietVsSignalingNaN(TotalOrderMagFunc func) {
+    EXPECT_FALSE(funcWrapper(func, neg_aNaN, neg_sNaN));
+    EXPECT_TRUE(funcWrapper(func, neg_sNaN, neg_aNaN));
+    EXPECT_TRUE(funcWrapper(func, sNaN, aNaN));
+    EXPECT_FALSE(funcWrapper(func, aNaN, sNaN));
+  }
+
+  void testNaNPayloads(TotalOrderMagFunc func) {
+    T qnan_123 = FPBits::quiet_nan(Sign::POS, 0x123).get_val();
+    T neg_qnan_123 = FPBits::quiet_nan(Sign::NEG, 0x123).get_val();
+    T snan_123 = FPBits::signaling_nan(Sign::POS, 0x123).get_val();
+    T neg_snan_123 = FPBits::signaling_nan(Sign::NEG, 0x123).get_val();
+
+    EXPECT_TRUE(funcWrapper(func, aNaN, aNaN));
+    EXPECT_TRUE(funcWrapper(func, neg_aNaN, neg_aNaN));
+    EXPECT_TRUE(funcWrapper(func, sNaN, sNaN));
+    EXPECT_TRUE(funcWrapper(func, neg_sNaN, neg_sNaN));
+
+    EXPECT_TRUE(funcWrapper(func, aNaN, qnan_123));
+    EXPECT_TRUE(funcWrapper(func, neg_aNaN, neg_qnan_123));
+    EXPECT_TRUE(funcWrapper(func, sNaN, snan_123));
+    EXPECT_TRUE(funcWrapper(func, neg_sNaN, neg_snan_123));
+
+    EXPECT_FALSE(funcWrapper(func, qnan_123, aNaN));
+    EXPECT_FALSE(funcWrapper(func, neg_qnan_123, neg_aNaN));
+    EXPECT_FALSE(funcWrapper(func, snan_123, sNaN));
+    EXPECT_FALSE(funcWrapper(func, neg_snan_123, neg_sNaN));
+  }
+};
+
+#define LIST_TOTALORDERMAG_TESTS(T, func)                                      \
+  using LlvmLibcTotalOrderMagTest = TotalOrderMagTestTemplate<T>;              \
+  TEST_F(LlvmLibcTotalOrderMagTest, XLesserThanY) { testXLesserThanY(&func); } \
+  TEST_F(LlvmLibcTotalOrderMagTest, XGreaterThanY) {                           \
+    testXGreaterThanY(&func);                                                  \
+  }                                                                            \
+  TEST_F(LlvmLibcTotalOrderMagTest, XEqualToY) { testXEqualToY(&func); }       \
+  TEST_F(LlvmLibcTotalOrderMagTest, SingleNaN) { testSingleNaN(&func); }       \
+  TEST_F(LlvmLibcTotalOrderMagTest, NaNSigns) { testNaNSigns(&func); }         \
+  TEST_F(LlvmLibcTotalOrderMagTest, QuietVsSignalingNaN) {                     \
+    testQuietVsSignalingNaN(&func);                                            \
+  }                                                                            \
+  TEST_F(LlvmLibcTotalOrderMagTest, NaNPayloads) { testNaNPayloads(&func); }
+
+#endif // LIBC_TEST_SRC_MATH_SMOKE_TOTALORDERMAGTEST_H
diff --git a/libc/test/src/math/smoke/TotalOrderTest.h b/libc/test/src/math/smoke/TotalOrderTest.h
n...
[truncated]

@overmighty overmighty mentioned this pull request Jun 10, 2024
Closed
70 tasks
overmighty
overmighty commented Jun 10, 2024
View reviewed changes
libc/src/__support/FPUtil/BasicOperations.h Outdated
Comment on lines 278 to 280
// Otherwise, the order of NaNs is implementation-defined (IEEE 754-2019).
// We order by payload.
return xbits.get_nan_payload() <= ybits.get_nan_payload();
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess this should be documented in https://libc.llvm.org/dev/undefined_behavior.html.

libc/test/src/math/smoke/TotalOrderMagTest.h
Comment on lines +107 to +110
T qnan_123 = FPBits::quiet_nan(Sign::POS, 0x123).get_val();
T neg_qnan_123 = FPBits::quiet_nan(Sign::NEG, 0x123).get_val();
T snan_123 = FPBits::signaling_nan(Sign::POS, 0x123).get_val();
T neg_snan_123 = FPBits::signaling_nan(Sign::NEG, 0x123).get_val();
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't really know if I'm supposed to keep these variable names in snake_case, including the nan part, or write NaN instead.

Also, why do the tests use camelCase for member function names?

libc/test/src/math/smoke/TotalOrderMagTest.h
@@ -0,0 +1,143 @@
//===-- Utility class to test flavors of totalordermag ----------*- C++ -*-===//
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I omitted the word "different" in "test [different] flavors", unlike usual, so that -*- C++ -*- can fit.

@lntue lntue merged commit f5dcfb9 into llvm:main Jun 11, 2024
5 of 6 checks passed
@HerrCai0907 HerrCai0907 mentioned this pull request Jun 13, 2024
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@lntue lntue lntue approved these changes

Assignees
No one assigned
Labels
libc
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@overmighty @llvmbot @lntue