[libc][math][c23] Add rsqrtf() function #159615

amemov · 2025-09-18T17:46:01Z

Changes:

Initial implementation of rsqrt for single precision float

Some small unrelated style changes to this PR (that I missed in my rsqrtf16 PR):

Added extra - to the top comments to make it look nicer in libc/shared/math/rsqrtf16.h
Put rsqrtf16 inside of libc/src/__support/math/CMakeLists.txt in sorted order
Rearanged libc_math_function rsqrtf16 in Bazel to match alphabetical order

Some small unrelated changes to this PR: - Added extra - to the top comments to make it look nicer in libc/shared/math/rsqrtf16.h - Put rsqrtf16 inside of libc/src/__support/math/CMakeLists.txt in sorted order - Rearanged libc_math_function rsqrtf16 in Bazel to match alphabetical order

llvmbot · 2025-09-18T17:46:33Z

@llvm/pr-subscribers-backend-risc-v

@llvm/pr-subscribers-libc

Author: Anton Shepelev (amemov)

Changes

Addresses #159614

Changes:

Initial implementation of rsqrt for single precision float

Some small unrelated style changes to this PR (that I missed in my rsqrtf16 PR):

Added extra - to the top comments to make it look nicer in libc/shared/math/rsqrtf16.h
Put rsqrtf16 inside of libc/src/__support/math/CMakeLists.txt in sorted order
Rearanged libc_math_function rsqrtf16 in Bazel to match alphabetical order

Full diff: https://github.com/llvm/llvm-project/pull/159615.diff

20 Files Affected:

(modified) libc/config/linux/aarch64/entrypoints.txt (+1)
(modified) libc/config/linux/riscv/entrypoints.txt (+1)
(modified) libc/config/linux/x86_64/entrypoints.txt (+1)
(modified) libc/docs/headers/math/index.rst (+1-1)
(modified) libc/include/math.yaml (+6)
(modified) libc/shared/math.h (+1-1)
(added) libc/shared/math/rsqrtf.h (+23)
(modified) libc/shared/math/rsqrtf16.h (+1-1)
(modified) libc/src/__support/math/CMakeLists.txt (+30-16)
(modified) libc/src/math/CMakeLists.txt (+1)
(modified) libc/src/math/generic/CMakeLists.txt (+11)
(added) libc/src/math/generic/rsqrtf.cpp (+15)
(added) libc/src/math/rsqrtf.h (+20)
(modified) libc/test/shared/CMakeLists.txt (+1-1)
(modified) libc/test/shared/shared_math_test.cpp (+1)
(modified) libc/test/src/math/CMakeLists.txt (+11)
(added) libc/test/src/math/rsqrtf_test.cpp (+41)
(modified) libc/test/src/math/smoke/CMakeLists.txt (+11)
(added) libc/test/src/math/smoke/rsqrtf_test.cpp (+42)
(modified) utils/bazel/llvm-project-overlay/libc/BUILD.bazel (+30-8)

diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt
index 00c4b2ec0f828..9e04eedd16162 100644
--- a/libc/config/linux/aarch64/entrypoints.txt
+++ b/libc/config/linux/aarch64/entrypoints.txt
@@ -735,6 +735,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
     libc.src.math.rintf16
     libc.src.math.roundevenf16
     libc.src.math.roundf16
+    libc.src.math.rsqrtf
     libc.src.math.rsqrtf16
     libc.src.math.scalblnf16
     libc.src.math.scalbnf16
diff --git a/libc/config/linux/riscv/entrypoints.txt b/libc/config/linux/riscv/entrypoints.txt
index 89e3653186d13..f5742faf5355e 100644
--- a/libc/config/linux/riscv/entrypoints.txt
+++ b/libc/config/linux/riscv/entrypoints.txt
@@ -749,6 +749,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
     libc.src.math.rintf16
     libc.src.math.roundevenf16
     libc.src.math.roundf16
+    libc.src.math.rsqrtf
     libc.src.math.rsqrtf16
     libc.src.math.scalblnf16
     libc.src.math.scalbnf16
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index 0bb8a683c5b01..0b861c537df9e 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -784,6 +784,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
     libc.src.math.rintf16
     libc.src.math.roundevenf16
     libc.src.math.roundf16
+    libc.src.math.rsqrtf
     libc.src.math.rsqrtf16
     libc.src.math.scalblnf16
     libc.src.math.scalbnf16
diff --git a/libc/docs/headers/math/index.rst b/libc/docs/headers/math/index.rst
index 7d5b341ba674a..392b419b2d949 100644
--- a/libc/docs/headers/math/index.rst
+++ b/libc/docs/headers/math/index.rst
@@ -343,7 +343,7 @@ Higher Math Functions
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+------------------------+----------------------------+
 | rootn     |                  |                 |                        |                      |                        |                        | 7.12.7.8               | F.10.4.8                   |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+------------------------+----------------------------+
-| rsqrt     |                  |                 |                        | |check|              |                        |                        | 7.12.7.9               | F.10.4.9                   |
+| rsqrt     | |check|          |                 |                        | |check|              |                        |                        | 7.12.7.9               | F.10.4.9                   |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+------------------------+----------------------------+
 | sin       | |check|          | |check|         |                        | |check|              |                        |                        | 7.12.4.6               | F.10.1.6                   |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+------------------------+----------------------------+
diff --git a/libc/include/math.yaml b/libc/include/math.yaml
index 6c800a0e2aa28..afd3ae33305c1 100644
--- a/libc/include/math.yaml
+++ b/libc/include/math.yaml
@@ -2349,6 +2349,12 @@ functions:
     return_type: long double
     arguments:
       - type: long double
+  - name: rsqrtf
+    standards:
+      - stdc
+    return_type: float
+    arguments:
+      - type: float
   - name: rsqrtf16
     standards:
       - stdc
diff --git a/libc/shared/math.h b/libc/shared/math.h
index 4f20095912bf1..abcea786a28b5 100644
--- a/libc/shared/math.h
+++ b/libc/shared/math.h
@@ -52,7 +52,7 @@
 #include "math/ldexpf.h"
 #include "math/ldexpf128.h"
 #include "math/ldexpf16.h"
-
+#include "math/rsqrtf.h"
 #include "math/rsqrtf16.h"
 
 #endif // LLVM_LIBC_SHARED_MATH_H
diff --git a/libc/shared/math/rsqrtf.h b/libc/shared/math/rsqrtf.h
new file mode 100644
index 0000000000000..32d0a61f7c21b
--- /dev/null
+++ b/libc/shared/math/rsqrtf.h
@@ -0,0 +1,23 @@
+//===-- Shared rsqrtf function ----------------------------------*- 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_RSQRTF_H
+#define LLVM_LIBC_SHARED_MATH_RSQRTF_H
+
+#include "shared/libc_common.h"
+#include "src/__support/math/rsqrtf.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace shared {
+
+using math::rsqrtf;
+
+} // namespace shared
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SHARED_MATH_RSQRTF_H
diff --git a/libc/shared/math/rsqrtf16.h b/libc/shared/math/rsqrtf16.h
index 54c7499214636..8dbf065802f49 100644
--- a/libc/shared/math/rsqrtf16.h
+++ b/libc/shared/math/rsqrtf16.h
@@ -1,4 +1,4 @@
-//===-- Shared rsqrtf16 function -------------------------------*- C++ -*-===//
+//===-- Shared rsqrtf16 function --------------------------------*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
diff --git a/libc/src/__support/math/CMakeLists.txt b/libc/src/__support/math/CMakeLists.txt
index ed5f314b0a9b5..26ad5a1d8901a 100644
--- a/libc/src/__support/math/CMakeLists.txt
+++ b/libc/src/__support/math/CMakeLists.txt
@@ -109,22 +109,6 @@ add_header_library(
     libc.src.__support.macros.properties.types
 )
 
-
-add_header_library(
-  rsqrtf16
-  HDRS
-    rsqrtf16.h
-  DEPENDS
-    libc.src.__support.FPUtil.cast
-    libc.src.__support.FPUtil.fenv_impl
-    libc.src.__support.FPUtil.fp_bits
-    libc.src.__support.FPUtil.multiply_add
-    libc.src.__support.FPUtil.polyeval
-    libc.src.__support.FPUtil.manipulation_functions
-    libc.src.__support.macros.optimization
-    libc.src.__support.macros.properties.types
-)
-
 add_header_library(
   asin_utils
   HDRS
@@ -775,6 +759,36 @@ add_header_library(
     libc.src.__support.common
 )
 
+add_header_library(
+  rsqrtf16
+  HDRS
+    rsqrtf16.h
+  DEPENDS
+    libc.src.__support.FPUtil.cast
+    libc.src.__support.FPUtil.fenv_impl
+    libc.src.__support.FPUtil.fp_bits
+    libc.src.__support.FPUtil.multiply_add
+    libc.src.__support.FPUtil.polyeval
+    libc.src.__support.FPUtil.manipulation_functions
+    libc.src.__support.macros.optimization
+    libc.src.__support.macros.properties.types
+)
+
+add_header_library(
+  rsqrtf
+  HDRS
+    rsqrtf.h
+  DEPENDS
+    libc.src.__support.FPUtil.cast
+    libc.src.__support.FPUtil.fenv_impl
+    libc.src.__support.FPUtil.fp_bits
+    libc.src.__support.FPUtil.multiply_add
+    libc.src.__support.FPUtil.polyeval
+    libc.src.__support.FPUtil.manipulation_functions
+    libc.src.__support.macros.optimization
+    libc.src.__support.macros.properties.types
+)
+
 add_header_library(
   sincos_eval
   HDRS
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index a6f400c873b7e..271c1c0ae1045 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -516,6 +516,7 @@ add_math_entrypoint_object(roundevenf16)
 add_math_entrypoint_object(roundevenf128)
 add_math_entrypoint_object(roundevenbf16)
 
+add_math_entrypoint_object(rsqrtf)
 add_math_entrypoint_object(rsqrtf16)
 
 add_math_entrypoint_object(scalbln)
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index ca7baeccae01a..2d92d7342ec96 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -988,6 +988,17 @@ add_entrypoint_object(
     ROUND_OPT
 )
 
+add_entrypoint_object(
+  rsqrtf
+  SRCS
+    rsqrtf.cpp
+  HDRS
+    ../rsqrtf.h
+  DEPENDS
+    libc.src.__support.math.rsqrtf
+    libc.src.errno.errno
+)
+
 add_entrypoint_object(
   rsqrtf16
   SRCS
diff --git a/libc/src/math/generic/rsqrtf.cpp b/libc/src/math/generic/rsqrtf.cpp
new file mode 100644
index 0000000000000..1c37979b35ed5
--- /dev/null
+++ b/libc/src/math/generic/rsqrtf.cpp
@@ -0,0 +1,15 @@
+//===-- Single-precision rsqrt 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/rsqrtf.h"
+#include "src/__support/math/rsqrtf.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(float, rsqrtf, (float x)) { return math::rsqrtf(x); }
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/math/rsqrtf.h b/libc/src/math/rsqrtf.h
new file mode 100644
index 0000000000000..bc559043be828
--- /dev/null
+++ b/libc/src/math/rsqrtf.h
@@ -0,0 +1,20 @@
+//===-- Implementation header for rsqrtf ------------------------*- 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_RSQRTF_H
+#define LLVM_LIBC_SRC_MATH_RSQRTF_H
+
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+float rsqrtf(float x);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_MATH_RSQRTF_H
diff --git a/libc/test/shared/CMakeLists.txt b/libc/test/shared/CMakeLists.txt
index 495d6f0a81a4c..0fa89e3ae6b62 100644
--- a/libc/test/shared/CMakeLists.txt
+++ b/libc/test/shared/CMakeLists.txt
@@ -48,6 +48,6 @@ add_fp_unittest(
     libc.src.__support.math.ldexpf
     libc.src.__support.math.ldexpf128
     libc.src.__support.math.ldexpf16
+    libc.src.__support.math.rsqrtf
     libc.src.__support.math.rsqrtf16
-
 )
diff --git a/libc/test/shared/shared_math_test.cpp b/libc/test/shared/shared_math_test.cpp
index aa459f88c29f5..46bd112549cbd 100644
--- a/libc/test/shared/shared_math_test.cpp
+++ b/libc/test/shared/shared_math_test.cpp
@@ -67,6 +67,7 @@ TEST(LlvmLibcSharedMathTest, AllFloat) {
 
   ASSERT_FP_EQ(float(8 << 5), LIBC_NAMESPACE::shared::ldexpf(8.0f, 5));
   ASSERT_FP_EQ(float(-1 * (8 << 5)), LIBC_NAMESPACE::shared::ldexpf(-8.0f, 5));
+  EXPECT_FP_EQ(0x1p+0f, LIBC_NAMESPACE::shared::rsqrtf(1.0f));
 }
 
 TEST(LlvmLibcSharedMathTest, AllDouble) {
diff --git a/libc/test/src/math/CMakeLists.txt b/libc/test/src/math/CMakeLists.txt
index 9d644703a61ae..f728eb83ebcf4 100644
--- a/libc/test/src/math/CMakeLists.txt
+++ b/libc/test/src/math/CMakeLists.txt
@@ -1678,6 +1678,17 @@ add_fp_unittest(
     libc.src.math.sqrtl
 )
 
+add_fp_unittest(
+  rsqrtf_test
+  NEED_MPFR
+  SUITE
+    libc-math-unittests
+  SRCS
+    rsqrtf_test.cpp
+  DEPENDS
+    libc.src.math.rsqrtf
+)
+
 add_fp_unittest(
   rsqrtf16_test
   NEED_MPFR
diff --git a/libc/test/src/math/rsqrtf_test.cpp b/libc/test/src/math/rsqrtf_test.cpp
new file mode 100644
index 0000000000000..cedafb63a31de
--- /dev/null
+++ b/libc/test/src/math/rsqrtf_test.cpp
@@ -0,0 +1,41 @@
+//===-- Exhaustive test for rsqrtf ----------------------------------------===//
+//
+// 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/rsqrtf.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+#include "utils/MPFRWrapper/MPFRUtils.h"
+
+using LlvmLibcRsqrtfTest = LIBC_NAMESPACE::testing::FPTest<float>;
+
+namespace mpfr = LIBC_NAMESPACE::testing::mpfr;
+
+// Range: [0, Inf]
+static constexpr uint32_t POS_START = 0x00000000u; 
+static constexpr uint32_t POS_STOP = 0x7F800000u;
+
+// Range: [-Inf, 0)
+// rsqrt(-0.0) is -inf, not the same for mpfr.
+static constexpr uint32_t NEG_START = 0x80000001u;
+static constexpr uint32_t NEG_STOP = 0xFF800000u;
+
+TEST_F(LlvmLibcRsqrtfTest, PositiveRange) {
+  for (uint32_t v = POS_START; v <= POS_STOP; ++v) {
+    float x = FPBits(v).get_val();
+    EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Rsqrt, x,
+                                   LIBC_NAMESPACE::rsqrtf(x), 0.5);
+  }
+}
+
+TEST_F(LlvmLibcRsqrtfTest, NegativeRange) {
+  for (uint32_t v = NEG_START; v <= NEG_STOP; ++v) {
+    float x = FPBits(v).get_val();
+    EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Rsqrt, x,
+                                   LIBC_NAMESPACE::rsqrtf(x), 0.5);
+  }
+}
diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt
index eadd5f0970722..644fa90a412fa 100644
--- a/libc/test/src/math/smoke/CMakeLists.txt
+++ b/libc/test/src/math/smoke/CMakeLists.txt
@@ -3502,6 +3502,17 @@ add_fp_unittest(
     libc.src.math.sqrtl
 )
 
+add_fp_unittest(
+  rsqrtf_test
+  SUITE
+    libc-math-smoke-tests
+  SRCS
+    rsqrtf_test.cpp
+  DEPENDS
+    libc.src.math.rsqrtf
+    libc.hdr.errno_macros
+)
+
 add_fp_unittest(
   rsqrtf16_test
   SUITE
diff --git a/libc/test/src/math/smoke/rsqrtf_test.cpp b/libc/test/src/math/smoke/rsqrtf_test.cpp
new file mode 100644
index 0000000000000..9ffde79e5e9a7
--- /dev/null
+++ b/libc/test/src/math/smoke/rsqrtf_test.cpp
@@ -0,0 +1,42 @@
+//===-- Unittests for rsqrtf --------------------------------------------===//
+//
+// 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 "hdr/errno_macros.h"
+#include "src/math/rsqrtf.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+
+using LlvmLibcRsqrtfTest = LIBC_NAMESPACE::testing::FPTest<float>;
+
+TEST_F(LlvmLibcRsqrtfTest, SpecialNumbers) {
+  EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::rsqrtf(aNaN));
+  EXPECT_MATH_ERRNO(0);
+
+  EXPECT_FP_IS_NAN_WITH_EXCEPTION(LIBC_NAMESPACE::rsqrtf(sNaN), FE_INVALID);
+  EXPECT_MATH_ERRNO(0);
+
+  EXPECT_FP_EQ(inf, LIBC_NAMESPACE::rsqrtf(zero));
+  EXPECT_MATH_ERRNO(ERANGE);
+
+  EXPECT_FP_EQ(neg_inf, LIBC_NAMESPACE::rsqrtf(neg_zero));
+  EXPECT_MATH_ERRNO(ERANGE);
+
+  EXPECT_FP_EQ(
+      1.0f,
+      LIBC_NAMESPACE::rsqrtf(1.0f));
+  EXPECT_MATH_ERRNO(0);
+
+  EXPECT_FP_EQ(zero, LIBC_NAMESPACE::rsqrtf(inf));
+  EXPECT_MATH_ERRNO(0);
+
+  EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::rsqrtf(neg_inf));
+  EXPECT_MATH_ERRNO(EDOM);
+
+  EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::rsqrtf(-2.0f));
+  EXPECT_MATH_ERRNO(EDOM);
+}
diff --git a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
index 921163a413067..87ad0926bfee6 100644
--- a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
+++ b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
@@ -2296,6 +2296,20 @@ libc_support_library(
     ],
 )
 
+libc_support_library(
+    name = "__support_math_rsqrtf",
+    hdrs = ["src/__support/math/rsqrtf.h"],
+    deps = [
+        ":__support_fputil_fenv_impl",
+        ":__support_fputil_fp_bits",
+        ":__support_fputil_manipulation_functions",
+        ":__support_fputil_multiply_add",
+        ":__support_fputil_polyeval",
+        ":__support_macros_optimization",
+        ":__support_macros_properties_types",
+    ],
+)
+
 libc_support_library(
     name = "__support_math_rsqrtf16",
     hdrs = ["src/__support/math/rsqrtf16.h"],
@@ -3259,14 +3273,6 @@ libc_math_function(
     ],
 )
 
-libc_math_function(
-    name = "rsqrtf16",
-    additional_deps = [
-        ":__support_math_rsqrtf16",
-        ":errno",
-    ],
-)
-
 libc_math_function(
     name = "acoshf16",
     additional_deps = [
@@ -4497,6 +4503,22 @@ libc_math_function(name = "roundevenf128")
 
 libc_math_function(name = "roundevenf16")
 
+libc_math_function(
+    name = "rsqrtf",
+    additional_deps = [
+        ":__support_math_rsqrtf",
+        ":errno",
+    ],
+)
+
+libc_math_function(
+    name = "rsqrtf16",
+    additional_deps = [
+        ":__support_math_rsqrtf16",
+        ":errno",
+    ],
+)
+
 libc_math_function(name = "scalbln")
 
 libc_math_function(name = "scalblnf")

github-actions · 2025-09-18T17:49:46Z

✅ With the latest revision this PR passed the C/C++ code formatter.

- added missing header

amemov · 2025-09-18T19:03:59Z

@lntue
I tried here to use just float for 1/sqrt() and then return it, but it turned out that when I was running the exhaustive tests (which I need to change so it wouldn't take forever to run ninja check-libc) it would fail all the time - which make sense. At least when I left the test running on my computer just for fun overnight it didn't show any errors (but at the same time it didn't finish the test and I had to stop it).

I implemented the rsqrtf here similarly to how I did rsqrtf16, but instead of using float, I used double and casted it to float. I suppose this implementation would be ok for now, and like with rsqrtf16 in the future we would need to change it to use approximation -- in this particular case, if hardware doesn't have double precision, need to fallback to float-based approximation

- Created unit test template for rsqrt functions that can be called in the future by other precision functions - Changed range of values for unit-test to match the libc/test/src/math/SqrtTest.h - Changed the comments in rsqrtf.h, removed unnecessary code in the comments

overmighty · 2025-09-24T18:24:41Z

@amemov FYI, you need to write "Fixes #(issue number)" or "Closes #(issue number)" (or a few variations of these) instead of "Addresses #(issue number)" for GitHub to automatically close the issue once the PR is merged.

libc/src/__support/math/CMakeLists.txt

libc/src/__support/math/rsqrtf.h

libc/test/src/math/exhaustive/CMakeLists.txt

libc/test/src/math/smoke/rsqrtf_test.cpp

libc/test/src/math/smoke/CMakeLists.txt

utils/bazel/llvm-project-overlay/libc/BUILD.bazel

…to match rsqrtf_test.cpp

overmighty · 2025-10-04T12:00:54Z

libc/src/__support/math/CMakeLists.txt

+    libc.hdr.errno_macros  
+    libc.hdr.fenv_macros  
+    libc.src.__support.FPUtil.cast  
+    libc.src.__support.FPUtil.fenv_impl  
+    libc.src.__support.FPUtil.fp_bits  
+    libc.src.__support.macros.optimization


Nit: trailing whitespace (I wish GitHub would highlight it.).

Missing libc.src.__support.FPUtil.sqrt.

Suggested change

libc.hdr.errno_macros

libc.hdr.fenv_macros

libc.src.__support.FPUtil.cast

libc.src.__support.FPUtil.fenv_impl

libc.src.__support.FPUtil.fp_bits

libc.src.__support.macros.optimization

libc.hdr.errno_macros

libc.hdr.fenv_macros

libc.src.__support.FPUtil.cast

libc.src.__support.FPUtil.fenv_impl

libc.src.__support.FPUtil.fp_bits

libc.src.__support.FPUtil.sqrt

libc.src.__support.macros.optimization

overmighty · 2025-10-04T12:05:36Z

libc/test/shared/shared_math_test.cpp


  ASSERT_FP_EQ(float(8 << 5), LIBC_NAMESPACE::shared::ldexpf(8.0f, 5));
  ASSERT_FP_EQ(float(-1 * (8 << 5)), LIBC_NAMESPACE::shared::ldexpf(-8.0f, 5));
+  EXPECT_FP_EQ(0x1p+0f, LIBC_NAMESPACE::shared::rsqrtf(1.0f));


Oh, I didn't notice the alphabetical order. It just looks kind of weird to me to have 2 test cases for ldexpf followed by a test case for rsqrtf with no blank line in between because it makes it look like these tests are related somehow. Lines 49-62 don't look weird to me because there's only 1 test case per function. But this is an extreme nitpick.

overmighty · 2025-10-04T12:07:59Z

utils/bazel/llvm-project-overlay/libc/BUILD.bazel

+    deps = [
+        ":__support_fputil_cast",
+        ":__support_fputil_fenv_impl",
+        ":__support_fputil_fp_bits",


Suggested change

":__support_fputil_fp_bits",

":__support_fputil_fp_bits",

":__support_fputil_sqrt",

Why did you remove cast? If I understand correctly, here in the bazel it should add the same dependencies as CMake in libc/src/__support/math , which are:

libc.hdr.errno_macros libc.hdr.fenv_macros libc.src.__support.FPUtil.cast libc.src.__support.FPUtil.fenv_impl libc.src.__support.FPUtil.fp_bits libc.src.__support.FPUtil.sqrt libc.src.__support.macros.optimization

overmighty · 2025-10-04T12:09:36Z

utils/bazel/llvm-project-overlay/libc/BUILD.bazel

+libc_math_function(
+    name = "rsqrtf",
+    additional_deps = [
+        ":__support_math_rsqrtf",
+    ],
+)
+
+libc_math_function(
+    name = "rsqrtf16",
+    additional_deps = [
+        ":__support_math_rsqrtf16",
+        ":errno",
+    ],
+)


Why does rsqrtf16 depend on :errno?

Will remove that

amemov requested review from rupprecht, keith and aaronmondal as code owners September 18, 2025 17:46

amemov marked this pull request as draft September 18, 2025 17:46

llvmbot added backend:RISC-V libc bazel "Peripheral" support tier build system: utils/bazel labels Sep 18, 2025

- clang-formatted the files

5b8b21a

- added missing header

amemov marked this pull request as ready for review September 19, 2025 19:05

- Added comments to RsqrtTest.h template, to clarify use case

0d8e57a

lntue requested review from overmighty, krishna2803 and lntue September 19, 2025 20:54

overmighty requested changes Sep 24, 2025

View reviewed changes

amemov added 2 commits September 24, 2025 13:44

Addressed the comments, fixed include in rsqrtf16_test.cpp and cmake …

1cc47f3

…to match rsqrtf_test.cpp

Clang-formatted the files in question, ran Buildifier on .bazel

0faacb6

amemov requested a review from overmighty September 24, 2025 21:35

overmighty requested changes Oct 4, 2025

View reviewed changes

overmighty reviewed Oct 4, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[libc][math][c23] Add rsqrtf() function #159615

[libc][math][c23] Add rsqrtf() function #159615

amemov commented Sep 18, 2025 •

edited

Loading

Uh oh!

llvmbot commented Sep 18, 2025 •

edited

Loading

Uh oh!

github-actions bot commented Sep 18, 2025 •

edited

Loading

Uh oh!

amemov commented Sep 18, 2025 •

edited

Loading

Uh oh!

overmighty commented Sep 24, 2025 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

overmighty Oct 4, 2025

Uh oh!

overmighty Oct 4, 2025

Uh oh!

overmighty Oct 4, 2025

Uh oh!

amemov Oct 4, 2025

Uh oh!

overmighty Oct 4, 2025

Uh oh!

amemov Oct 4, 2025

Uh oh!

Uh oh!

	":__support_fputil_fp_bits",
	":__support_fputil_fp_bits",
	":__support_fputil_sqrt",

[libc][math][c23] Add rsqrtf() function #159615

Are you sure you want to change the base?

[libc][math][c23] Add rsqrtf() function #159615

Conversation

amemov commented Sep 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Sep 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Sep 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

amemov commented Sep 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

overmighty commented Sep 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

overmighty Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

overmighty Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

overmighty Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

amemov Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

overmighty Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

amemov Oct 4, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

amemov commented Sep 18, 2025 •

edited

Loading

llvmbot commented Sep 18, 2025 •

edited

Loading

github-actions bot commented Sep 18, 2025 •

edited

Loading

amemov commented Sep 18, 2025 •

edited

Loading

overmighty commented Sep 24, 2025 •

edited

Loading