[libc][math][c23] implement C23 `acospif` math function by hulxv · Pull Request #183661 · llvm/llvm-project

hulxv · 2026-02-27T01:47:47Z

Implementing C23 acospi math function for single-precision with the header-only approach that is followed since #147386

llvmbot · 2026-02-27T01:48:16Z

@llvm/pr-subscribers-libc

@llvm/pr-subscribers-backend-amdgpu

Author: Mohamed Emad (hulxv)

Changes

Implementing C23 acospi math function for single-precision with the header-only approach that is followed since #147386

Patch is 29.14 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/183661.diff

30 Files Affected:

(modified) libc/config/baremetal/aarch64/entrypoints.txt (+1)
(modified) libc/config/baremetal/arm/entrypoints.txt (+1)
(modified) libc/config/baremetal/riscv/entrypoints.txt (+1)
(modified) libc/config/darwin/aarch64/entrypoints.txt (+1)
(modified) libc/config/gpu/amdgpu/entrypoints.txt (+1)
(modified) libc/config/gpu/nvptx/entrypoints.txt (+1)
(modified) libc/config/linux/aarch64/entrypoints.txt (+1)
(modified) libc/config/linux/arm/entrypoints.txt (+1)
(modified) libc/config/linux/riscv/entrypoints.txt (+1)
(modified) libc/config/linux/x86_64/entrypoints.txt (+1)
(modified) libc/config/windows/entrypoints.txt (+1-1)
(added) libc/shared/math/acospif.h (+23)
(modified) libc/src/__support/math/CMakeLists.txt (+20-4)
(added) libc/src/__support/math/acospif.h (+114)
(modified) libc/src/__support/math/asinpif.h (+8-41)
(modified) libc/src/__support/math/inv_trigf_utils.h (+37)
(modified) libc/src/math/CMakeLists.txt (+1)
(added) libc/src/math/acospif.h (+20)
(modified) libc/src/math/generic/CMakeLists.txt (+12)
(added) libc/src/math/generic/acospif.cpp (+16)
(modified) libc/test/shared/CMakeLists.txt (+1)
(modified) libc/test/shared/shared_math_test.cpp (+1)
(modified) libc/test/src/math/CMakeLists.txt (+11)
(added) libc/test/src/math/acospif_test.cpp (+29)
(modified) libc/test/src/math/asinpif_test.cpp (+6-15)
(modified) libc/test/src/math/exhaustive/CMakeLists.txt (+16)
(added) libc/test/src/math/exhaustive/acospif_test.cpp (+33)
(modified) libc/test/src/math/smoke/CMakeLists.txt (+12)
(added) libc/test/src/math/smoke/acospif_test.cpp (+51)
(modified) utils/bazel/llvm-project-overlay/libc/BUILD.bazel (+24)

diff --git a/libc/config/baremetal/aarch64/entrypoints.txt b/libc/config/baremetal/aarch64/entrypoints.txt
index 582e7c52726fe..1c04ee2b6d040 100644
--- a/libc/config/baremetal/aarch64/entrypoints.txt
+++ b/libc/config/baremetal/aarch64/entrypoints.txt
@@ -336,6 +336,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/baremetal/arm/entrypoints.txt b/libc/config/baremetal/arm/entrypoints.txt
index 98374e6cede73..ad6b6d922b57a 100644
--- a/libc/config/baremetal/arm/entrypoints.txt
+++ b/libc/config/baremetal/arm/entrypoints.txt
@@ -339,6 +339,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/baremetal/riscv/entrypoints.txt b/libc/config/baremetal/riscv/entrypoints.txt
index b45b388ec9634..29be47dfbfd53 100644
--- a/libc/config/baremetal/riscv/entrypoints.txt
+++ b/libc/config/baremetal/riscv/entrypoints.txt
@@ -338,6 +338,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     # math.h entrypoints
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asinf
     libc.src.math.asinhf
     libc.src.math.asinpif
diff --git a/libc/config/darwin/aarch64/entrypoints.txt b/libc/config/darwin/aarch64/entrypoints.txt
index 5488bf1d0c7de..b4a1ee70764cb 100644
--- a/libc/config/darwin/aarch64/entrypoints.txt
+++ b/libc/config/darwin/aarch64/entrypoints.txt
@@ -153,6 +153,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/gpu/amdgpu/entrypoints.txt b/libc/config/gpu/amdgpu/entrypoints.txt
index 856bf1d55f3be..a65b6f0274fd8 100644
--- a/libc/config/gpu/amdgpu/entrypoints.txt
+++ b/libc/config/gpu/amdgpu/entrypoints.txt
@@ -283,6 +283,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/gpu/nvptx/entrypoints.txt b/libc/config/gpu/nvptx/entrypoints.txt
index 8765cf189f325..fee0038c88cc0 100644
--- a/libc/config/gpu/nvptx/entrypoints.txt
+++ b/libc/config/gpu/nvptx/entrypoints.txt
@@ -283,6 +283,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt
index d27dabaff5c19..960958b7cf90a 100644
--- a/libc/config/linux/aarch64/entrypoints.txt
+++ b/libc/config/linux/aarch64/entrypoints.txt
@@ -423,6 +423,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/linux/arm/entrypoints.txt b/libc/config/linux/arm/entrypoints.txt
index 714bd024dad1c..e76bf81670ce5 100644
--- a/libc/config/linux/arm/entrypoints.txt
+++ b/libc/config/linux/arm/entrypoints.txt
@@ -253,6 +253,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/linux/riscv/entrypoints.txt b/libc/config/linux/riscv/entrypoints.txt
index b823e1b25e6cc..a69d1a1b0a642 100644
--- a/libc/config/linux/riscv/entrypoints.txt
+++ b/libc/config/linux/riscv/entrypoints.txt
@@ -427,6 +427,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index f1b3e90cfd031..bb4973628d588 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -469,6 +469,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
     libc.src.math.asinhf
diff --git a/libc/config/windows/entrypoints.txt b/libc/config/windows/entrypoints.txt
index d88a49c9850ee..f777fc6c94bc2 100644
--- a/libc/config/windows/entrypoints.txt
+++ b/libc/config/windows/entrypoints.txt
@@ -133,9 +133,9 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acos
     libc.src.math.acosf
     libc.src.math.acoshf
+    libc.src.math.acospif
     libc.src.math.asin
     libc.src.math.asinf
-    libc.src.math.asinpif
     libc.src.math.asinhf
     libc.src.math.asinpif
     libc.src.math.atan2
diff --git a/libc/shared/math/acospif.h b/libc/shared/math/acospif.h
new file mode 100644
index 0000000000000..79a2ef697a85d
--- /dev/null
+++ b/libc/shared/math/acospif.h
@@ -0,0 +1,23 @@
+//===-- Shared acospif header -----------------------------------*- 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_SHARED_MATH_ACOSPIF_H
+#define LLVM_LIBC_SHARED_MATH_ACOSPIF_H
+
+#include "shared/libc_common.h"
+#include "src/__support/math/acospif.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace shared {
+
+using math::acospif;
+
+} // namespace shared
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SHARED_MATH_ACOSPIF_H
diff --git a/libc/src/__support/math/CMakeLists.txt b/libc/src/__support/math/CMakeLists.txt
index 75a55c2939e19..505ba8b80709f 100644
--- a/libc/src/__support/math/CMakeLists.txt
+++ b/libc/src/__support/math/CMakeLists.txt
@@ -94,6 +94,23 @@ add_header_library(
     libc.src.__support.macros.optimization
 )
 
+
+add_header_library(
+  acospif
+  HDRS
+    acospif.h
+  DEPENDS
+    .inv_trigf_utils
+    libc.src.__support.FPUtil.except_value_utils
+    libc.src.__support.FPUtil.fp_bits
+    libc.src.__support.FPUtil.multiply_add
+    libc.src.__support.FPUtil.polyeval
+    libc.src.__support.FPUtil.sqrt
+    libc.src.__support.macros.optimization
+)
+
+
+
 add_header_library(
   acospif16
   HDRS
@@ -177,12 +194,11 @@ add_header_library(
   HDRS
     asinpif.h
   DEPENDS
-    libc.src.__support.FPUtil.fenv_impl
-    libc.src.__support.FPUtil.fp_bits
-    libc.src.__support.FPUtil.polyeval
-    libc.src.__support.FPUtil.cast
+    .inv_trigf_utils
     libc.src.__support.FPUtil.except_value_utils
+    libc.src.__support.FPUtil.fp_bits
     libc.src.__support.FPUtil.multiply_add
+    libc.src.__support.FPUtil.polyeval
     libc.src.__support.FPUtil.sqrt
     libc.src.__support.macros.optimization
 )
diff --git a/libc/src/__support/math/acospif.h b/libc/src/__support/math/acospif.h
new file mode 100644
index 0000000000000..8fe0c3c232f45
--- /dev/null
+++ b/libc/src/__support/math/acospif.h
@@ -0,0 +1,114 @@
+//===-- Implementation header for acospif -----------------------*- 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___SUPPORT_MATH_ACOSPIF_H
+#define LLVM_LIBC_SRC___SUPPORT_MATH_ACOSPIF_H
+
+#include "inv_trigf_utils.h"
+#include "src/__support/FPUtil/FEnvImpl.h"
+#include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/PolyEval.h"
+#include "src/__support/FPUtil/cast.h"
+#include "src/__support/FPUtil/except_value_utils.h"
+#include "src/__support/FPUtil/multiply_add.h"
+#include "src/__support/FPUtil/sqrt.h"
+#include "src/__support/macros/optimization.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace math {
+
+LIBC_INLINE float acospif(float x) {
+#ifndef LIBC_MATH_HAS_SKIP_ACCURATE_PASS
+  constexpr size_t N_EXCEPTS = 0;
+  constexpr fputil::ExceptValues<float, N_EXCEPTS> ACOSPIF_EXCEPTS = {{
+      // (inputs, RZ output, RU offset, RD offset, RN offset)
+  }};
+#endif // !LIBC_MATH_HAS_SKIP_ACCURATE_PASS
+
+  using FPBits = fputil::FPBits<float>;
+
+  FPBits xbits(x);
+  bool is_neg = xbits.is_neg();
+  double x_abs = fputil::cast<double>(xbits.abs().get_val());
+
+  auto signed_result = [is_neg](auto r) -> auto { return is_neg ? -r : r; };
+
+  if (LIBC_UNLIKELY(x_abs > 1.0)) {
+    if (xbits.is_nan()) {
+      if (xbits.is_signaling_nan()) {
+        fputil::raise_except_if_required(FE_INVALID);
+        return FPBits::quiet_nan().get_val();
+      }
+      return x;
+    }
+
+    fputil::raise_except_if_required(FE_INVALID);
+    fputil::set_errno_if_required(EDOM);
+    return FPBits::quiet_nan().get_val();
+  }
+
+#ifndef LIBC_MATH_HAS_SKIP_ACCURATE_PASS
+  auto r = ACOSPIF_EXCEPTS.lookup(xbits.uintval());
+  if (LIBC_UNLIKELY(r.has_value()))
+    return r.value();
+#endif // !LIBC_MATH_HAS_SKIP_ACCURATE_PASS
+
+  // acospif(x) = 1/2 - asinpif(x)
+  //
+  // if |x| <= 0.5:
+  //   acospif(x) = 0.5 - x * (c0 + x^2 * P1(x^2))
+  if (LIBC_UNLIKELY(x_abs <= 0.5)) {
+    double x_d = fputil::cast<double>(x);
+    double v2 = x_d * x_d;
+    double result = x_d * fputil::multiply_add(
+                              v2, inv_trigf_utils_internal::asinpi_eval(v2),
+                              inv_trigf_utils_internal::ASINPI_COEFFS[0]);
+    return fputil::cast<float>(0.5 - result);
+  }
+
+  // If |x| > 0.5, we use the identity:
+  //   asinpif(x) = sign(x) * (0.5 - 2 * sqrt(u) * P(u))
+  // where u = (1 - |x|) / 2, P(u) ~ asin(sqrt(u)) / (pi * sqrt(u))
+  //
+  // Then:
+  //   acospif(x) = 0.5 - asinpif(x)
+  //
+  // For x > 0.5:
+  //   acospif(x) = 0.5 - (0.5 - 2*sqrt(u)*P(u)) = 2*sqrt(u)*P(u)
+  //
+  // For x < -0.5:
+  //   acospif(x) = 0.5 - (-(0.5 - 2*sqrt(u)*P(u))) = 1 - 2*sqrt(u)*P(u)
+
+  constexpr double ONE_OVER_PI_HI = 0x1.45f306dc9c883p-2;
+  constexpr double ONE_OVER_PI_LO = -0x1.6b01ec5417056p-56;
+  // C0_MINUS_1OVERPI = c0 - 1/pi = DELTA_C0 + ONE_OVER_PI_LO
+  constexpr double C0_MINUS_1OVERPI =
+      (inv_trigf_utils_internal::ASINPI_COEFFS[0] - ONE_OVER_PI_HI) +
+      ONE_OVER_PI_LO;
+
+  double u = fputil::multiply_add(-0.5, x_abs, 0.5);
+  double sqrt_u = fputil::sqrt<double>(u);
+  double neg2_sqrt_u = -2.0 * sqrt_u;
+
+  // tail = (c0 - 1/pi) + u * P1(u)
+  double tail = fputil::multiply_add(
+      u, inv_trigf_utils_internal::asinpi_eval(u), C0_MINUS_1OVERPI);
+
+  double result_hi = fputil::multiply_add(neg2_sqrt_u, ONE_OVER_PI_HI, 0.5);
+  double result = fputil::multiply_add(tail, neg2_sqrt_u, result_hi);
+
+  // For x > 0.5:  acospif(x) = 2*sqrt(u)*P(u)
+  // For x < -0.5: acospif(x) = 1 - 2*sqrt(u)*P(u)
+
+  return fputil::cast<float>(0.5 - signed_result(result));
+}
+
+} // namespace math
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC___SUPPORT_MATH_ACOSPIF_H
diff --git a/libc/src/__support/math/asinpif.h b/libc/src/__support/math/asinpif.h
index 398d64b4f131e..9a5daf6198a42 100644
--- a/libc/src/__support/math/asinpif.h
+++ b/libc/src/__support/math/asinpif.h
@@ -9,6 +9,7 @@
 #ifndef LLVM_LIBC_SRC___SUPPORT_MATH_ASINPIF_H
 #define LLVM_LIBC_SRC___SUPPORT_MATH_ASINPIF_H
 
+#include "inv_trigf_utils.h"
 #include "src/__support/FPUtil/FEnvImpl.h"
 #include "src/__support/FPUtil/FPBits.h"
 #include "src/__support/FPUtil/PolyEval.h"
@@ -66,50 +67,14 @@ LIBC_INLINE float asinpif(float x) {
     return r.value();
 #endif // !LIBC_MATH_HAS_SKIP_ACCURATE_PASS
 
-  // the coefficients for the polynomial approximation of asin(x)/(pi*x) in the
-  // range [0, 0.5] extracted using Sollya.
-  //
-  // Sollya code:
-  // > prec = 200;
-  // > display = hexadecimal;
-  // > g = asin(x) / (pi * x);
-  // > P = fpminimax(g, [|0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20|],
-  // >              [|D...|], [0, 0.5]);
-  // > for i from 0 to degree(P) do coeff(P, i);
-  // > print("Error:", dirtyinfnorm(P - g, [1e-30; 0.25]));
-  // Error: 0x1.45c281e1cf9b58p-50 ~= 2^−49.652
-  //
-  // Non-zero coefficients (even powers only):
-  constexpr double ASINPI_POLY_COEFFS[] = {
-      0x1.45f306dc9c881p-2,  // x^0
-      0x1.b2995e7b7e756p-5,  // x^2
-      0x1.8723a1d12f828p-6,  // x^4
-      0x1.d1a45564b9545p-7,  // x^6
-      0x1.3ce4ceaa0e1e9p-7,  // x^8
-      0x1.d2c305898ea13p-8,  // x^10
-      0x1.692212e27a5f9p-8,  // x^12
-      0x1.2b22cc744d25bp-8,  // x^14
-      0x1.8427b864479ffp-9,  // x^16
-      0x1.815522d7a2bf1p-8,  // x^18
-      -0x1.f6df98438aef4p-9, // x^20
-      0x1.4b50c2eb13708p-7   // x^22
-  };
-  // Evaluates P1(v2) = c1 + c2*v2 + c3*v2^2 + ... (tail of P without c0)
-  auto asinpi_polyeval = [&](double v2) -> double {
-    return fputil::polyeval(
-        v2, ASINPI_POLY_COEFFS[1], ASINPI_POLY_COEFFS[2], ASINPI_POLY_COEFFS[3],
-        ASINPI_POLY_COEFFS[4], ASINPI_POLY_COEFFS[5], ASINPI_POLY_COEFFS[6],
-        ASINPI_POLY_COEFFS[7], ASINPI_POLY_COEFFS[8], ASINPI_POLY_COEFFS[9],
-        ASINPI_POLY_COEFFS[10], ASINPI_POLY_COEFFS[11]);
-  };
-
   // if |x| <= 0.5:
   //   asinpi(x) = x * (c0 + x^2 * P1(x^2))
   if (LIBC_UNLIKELY(x_abs <= 0.5)) {
     double x_d = fputil::cast<double>(x);
     double v2 = x_d * x_d;
-    double result = x_d * fputil::multiply_add(v2, asinpi_polyeval(v2),
-                                               ASINPI_POLY_COEFFS[0]);
+    double result = x_d * fputil::multiply_add(
+                              v2, inv_trigf_utils_internal::asinpi_eval(v2),
+                              inv_trigf_utils_internal::ASINPI_COEFFS[0]);
     return fputil::cast<float>(result);
   }
 
@@ -131,14 +96,16 @@ LIBC_INLINE float asinpif(float x) {
   constexpr double ONE_OVER_PI_LO = -0x1.6b01ec5417056p-56;
   // C0_MINUS_1OVERPI = c0 - 1/pi = DELTA_C0 + ONE_OVER_PI_LO
   constexpr double C0_MINUS_1OVERPI =
-      (ASINPI_POLY_COEFFS[0] - ONE_OVER_PI_HI) + ONE_OVER_PI_LO;
+      (inv_trigf_utils_internal::ASINPI_COEFFS[0] - ONE_OVER_PI_HI) +
+      ONE_OVER_PI_LO;
 
   double u = fputil::multiply_add(-0.5, x_abs, 0.5);
   double sqrt_u = fputil::sqrt<double>(u);
   double neg2_sqrt_u = -2.0 * sqrt_u;
 
   // tail = (c0 - 1/pi) + u * P1(u)
-  double tail = fputil::multiply_add(u, asinpi_polyeval(u), C0_MINUS_1OVERPI);
+  double tail = fputil::multiply_add(
+      u, inv_trigf_utils_internal::asinpi_eval(u), C0_MINUS_1OVERPI);
 
   double result_hi = fputil::multiply_add(neg2_sqrt_u, ONE_OVER_PI_HI, 0.5);
   double result = fputil::multiply_add(tail, neg2_sqrt_u, result_hi);
diff --git a/libc/src/__support/math/inv_trigf_utils.h b/libc/src/__support/math/inv_trigf_utils.h
index 6eab68d8bbe01..7a93831333db0 100644
--- a/libc/src/__support/math/inv_trigf_utils.h
+++ b/libc/src/__support/math/inv_trigf_utils.h
@@ -177,6 +177,43 @@ LIBC_INLINE double asin_eval(double xsq) {
   return fputil::multiply_add(xsq, r2, r1);
 }
 
+// the coefficients for the polynomial approximation of asin(x)/(pi*x) in the
+// range [0, 0.5] extracted using Sollya.
+//
+// Sollya code:
+// > prec = 200;
+// > display = hexadecimal;
+// > g = asin(x) / (pi * x);
+// > P = fpminimax(g, [|0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20|],
+// >              [|D...|], [0, 0.5]);
+// > for i from 0 to degree(P) do coeff(P, i);
+// > print("Error:", dirtyinfnorm(P - g, [1e-30; 0.25]));
+// Error: 0x1.45c281e1cf9b58p-50 ~= 2^−49.652
+//
+// Non-zero coefficients (even powers only):
+LIBC_INLINE_VAR constexpr double ASINPI_COEFFS[13] = {
+    0x1.45f306dc9c881p-2,  // x^0
+    0x1.b2995e7b7e756p-5,  // x^2
+    0x1.8723a1d12f828p-6,  // x^4
+    0x1.d1a45564b9545p-7,  // x^6
+    0x1.3ce4ceaa0e1e9p-7,  // x^8
+    0x1.d2c305898ea13p-8,  // x^10
+    0x1.692212e27a5f9p-8,  // x^12
+    0x1.2b22cc744d25bp-8,  // x^14
+    0x1.8427b864479ffp-9,  // x^16
+    0x1.815522d7a2bf1p-8,  // x^18
+    -0x1.f6df98438aef4p-9, // x^20
+    0x1.4b50c2eb13708p-7   // x^22
+};
+
+// Evaluates P1(v2) = c1 + c2*v2 + c3*v2^2 + ... (tail of P without c0)
+LIBC_INLINE double asinpi_eval(double v2) {
+  return fputil::polyeval(
+      v2, ASINPI_COEFFS[1], ASINPI_COEFFS[2], ASINPI_COEFFS[3],
+      ASINPI_COEFFS[4], ASINPI_COEFFS[5], ASINPI_COEFFS[6], ASINPI_COEFFS[7],
+      ASINPI_COEFFS[8], ASINPI_COEFFS[9], ASINPI_COEFFS[10], ASINPI_COEFFS[11]);
+}
+
 } // namespace inv_trigf_utils_internal
 
 } // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index 7a9370fe487e2..e617950368994 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -48,6 +48,7 @@ add_math_entrypoint_object(acosh)
 add_math_entrypoint_object(acoshf)
 add_math_entrypoint_object(acoshf16)
 
+add_math_entrypoint_object(acospif)
 add_math_entrypoint_object(acospif16)
  
 add_math_entrypoint_object(asin)
diff --git a/libc/src/math/acospif.h b/libc/src/math/acospif.h
new file mode 100644
index 0000000000000..80e001a60d1ad
--- /dev/null
+++ b/libc/src/math/acospif.h
@@ -0,0 +1,20 @@
+//===-- Implementation header for acospif -----------------------*- 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_ACOSPIF_H
+#define LLVM_LIBC_SRC_MATH_ACOSPIF_H
+
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+float acospif(float x);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_MATH_ACOSPIF_H
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index f8834c5c12705..96e79833bad04 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -3985,6 +3985,18 @@ add_entrypoint_object(
     libc.src.errno.errno
 )
 
+add_entrypoint_object(
+  acospif
+  SRCS
+    acospif.cpp
+  HDRS
+    ../acospif.h
+  DEPENDS
+    libc.src.__support.math.acospif
+)
+
+
+
 add_entrypoint_object(
   asinhf
   SRCS
diff --git a/libc/src/math/generic/acospif.cpp b/libc/src/math/generic/acospif.cpp
new file mode 100644
index 0000000000000..928c2cba1441f
--- /dev/null
+++ b/libc/src/math/generic/acospif.cpp
@@ -0,0 +1,16 @@
+//===-- Single-precision acospif(x) 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/acospif.h"
+#include "src/__support/math/acospif.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(float, acospif, (float x)) { return math::acospif(x); }
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/test/shared/CMakeLists.txt b/libc/test/shared/CMakeLists.txt
index 490e498aaf9e8..79c0cb66a703c 100644
--- a/libc/test/shared/CMakeLists.txt
+++ b/libc/test/shared/CMakeLists.txt
@@ -13,6 +13,7 @@ add_fp_unittest(
     libc.src.__support.math.acosf16
     libc.src.__support.math.acoshf
     libc.src.__support.math.acoshf16
+    libc.src.__support.math.acospif
     libc.src.__support.math.acospif16
     libc.src.__support.math.asin
     libc.src.__support.math.asinf
diff --git a/libc/test/shared/shared_math_test.cpp b/libc/test/shared/shared_math_test.cpp
index 43ff7610e384b..7e607b5772616 100644
--- a/libc/test/shared/shared_math_test.cpp
+++ b/libc/test/shared/shared_math_test.cpp
@@ -146,6 +146,7 @@ TEST(LlvmLibcSharedMathTest, AllFloat) {
 
   EXPECT_FP_EQ(0x1.921fb6p+0, LIBC_NAMESPACE::shared::acosf(0.0f));
   EXPECT_FP_EQ(0x0p+0f, LIBC_NAMESPACE::shared::acoshf(1.0f));
+  EXPECT_FP_EQ(0x0p+0f, LIBC_NAMESPACE::shared::acospif(1.0f));
   EXPECT_FP_EQ(0x0p+0f, LIBC_NAMESPACE::shared::asinf(0.0f));
   EXPECT_FP_EQ(0x0p+0f, LIBC_NAMESPACE::shared:...
[truncated]

github-actions · 2026-02-27T01:54:41Z

🐧 Linux x64 Test Results

✅ The build succeeded and no tests ran. This is expected in some build configurations.

Implementing C23 `acospi` math function for single-precision with the header-only approach that is followed since llvm#147386

hulxv requested review from bassiounix and lntue February 27, 2026 01:47

hulxv requested review from aaronmondal, keith and rupprecht as code owners February 27, 2026 01:47

llvmbot added backend:AMDGPU backend:RISC-V libc bazel "Peripheral" support tier build system: utils/bazel labels Feb 27, 2026

hulxv force-pushed the libc/math/feat/acospif branch 2 times, most recently from 5fb453f to 13407ca Compare February 27, 2026 02:19

[libc][math][c23] implement C23 acospif math function

bd42e80

hulxv force-pushed the libc/math/feat/acospif branch from 13407ca to bd42e80 Compare February 27, 2026 03:40

hulxv mentioned this pull request Feb 27, 2026

[libc][math] Tracking Issue - Refactor Higher Math Functions to Header Only #147386

Closed

lntue reviewed Feb 27, 2026

View reviewed changes

Comment thread libc/src/__support/math/CMakeLists.txt Outdated

lntue reviewed Feb 27, 2026

View reviewed changes

Comment thread libc/src/math/generic/CMakeLists.txt Outdated

lntue approved these changes Feb 27, 2026

View reviewed changes

bassiounix requested changes Feb 27, 2026

View reviewed changes

Comment thread libc/src/__support/math/CMakeLists.txt

Comment thread libc/src/math/generic/CMakeLists.txt

nit

92df54b

hulxv requested a review from bassiounix February 27, 2026 15:47

bassiounix approved these changes Feb 28, 2026

View reviewed changes

hulxv merged commit 62cfe16 into llvm:main Feb 28, 2026
28 checks passed

sahas3 pushed a commit to sahas3/llvm-project that referenced this pull request Mar 4, 2026

[libc][math][c23] implement C23 acospif math function (llvm#183661)

6e972cc

Implementing C23 `acospi` math function for single-precision with the header-only approach that is followed since llvm#147386

This was referenced Apr 15, 2026

[X86] regenerate fcopysign test checks (#183710) ampandey-1995/llvm-project#4

Closed

[X86] regenerate fcopysign test checks (#183710) ampandey-1995/llvm-project#5

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[libc][math][c23] implement C23 `acospif` math function#183661

[libc][math][c23] implement C23 `acospif` math function#183661
hulxv merged 2 commits into
llvm:mainfrom
hulxv:libc/math/feat/acospif

hulxv commented Feb 27, 2026

Uh oh!

llvmbot commented Feb 27, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Feb 27, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

hulxv commented Feb 27, 2026

Uh oh!

llvmbot commented Feb 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Feb 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🐧 Linux x64 Test Results

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

llvmbot commented Feb 27, 2026 •

edited

Loading

github-actions Bot commented Feb 27, 2026 •

edited

Loading