Merge pull request #5875 from amyw-msft/patch-16

Update floating-point-support.md to call out OS version can impact results

Author: prmerger-automator[bot]

docs/c-runtime-library/floating-point-support.md

@@ -11,7 +11,9 @@ The Universal C Runtime library (UCRT) provides many integral and floating-point
 
 For ISO C Standard 11 (C11) and later, the `<tgmath.h>` header, in addition to including `<math.h>` and `<complex.h>`, provides macros that invoke a corresponding math function based on the types of the parameters. See [Type-generic math](tgmath.md) for details.
 
-Many of the floating point math library functions have different implementations for different CPU architectures. For example, the 32-bit x86 CRT may have a different implementation than the 64-bit x64 CRT. In addition, some of the functions may have multiple implementations for a given CPU architecture. The most efficient implementation is selected dynamically at run-time depending on the instruction sets supported by the CPU. For example, in the 32-bit x86 CRT, some functions have both an x87 implementation and an SSE2 implementation. When running on a CPU that supports SSE2, the faster SSE2 implementation is used. When running on a CPU that doesn't support SSE2, the slower x87 implementation is used. Because different implementations of the math library functions may use different CPU instructions and different algorithms to produce their results, the functions may produce different results across CPUs. In most cases, the results are within +/-1 ULP of the correctly rounded result, but the actual results may vary across CPUs.
+Many of the floating-point math library functions have different implementations for different CPU architectures. For example, the 32-bit x86 CRT may have a different implementation than the 64-bit x64 CRT. In addition, some of the functions may have multiple implementations for a given CPU architecture. The most efficient implementation is selected dynamically at run-time depending on the instruction sets supported by the CPU. For example, in the 32-bit x86 CRT, some functions have both an x87 implementation and an SSE2 implementation. When running on a CPU that supports SSE2, the faster SSE2 implementation is used. When running on a CPU that doesn't support SSE2, the slower x87 implementation is used. Because different implementations of the math library functions may use different CPU instructions and different algorithms to produce their results, the functions may produce different results across CPUs. In most cases, the results are within +/-1 ULP of the correctly rounded result, but the actual results may vary across CPUs.
+
+Newer versions of the UCRT might improve the precision and accuracy of the floating-point math library functions. Since the UCRT is part of the Windows operating system, you might get different results for these functions on different operating system versions or between debug and release builds. Although it is not recommended, you can statically link to the UCRT to guarantee consistent results if you need these functions will produce identical results everywhere.
 
 Previous 16-bit versions of Microsoft C/C++ and Microsoft Visual C++ supported the **`long double`** type as an 80-bit precision floating-point data type. In later versions of Visual C++, the **`long double`** data type is a 64-bit precision floating-point data type identical to the **`double`** type. The compiler treats **`long double`** and **`double`** as distinct types, but the **`long double`** functions are identical to their **`double`** counterparts. The CRT provides **`long double`** versions of the math functions for ISO C99 source code compatibility, but note that the binary representation may differ from other compilers.
 
@@ -79,7 +81,7 @@ Previous 16-bit versions of Microsoft C/C++ and Microsoft Visual C++ supported t
 | [`imaxabs`](./reference/imaxabs.md) | Computes the absolute value of an integer type |
 | [`imaxdiv`](./reference/imaxdiv.md) | Computes the quotient and the remainder of two integer values |
 | [`isfinite`, `_finite`, `_finitef`](./reference/finite-finitef.md) | Determines whether a value is finite |
-| [`isgreater`, `isgreaterequal`, `isless`, `islessequal`, `islessgreater`, `isunordered`](./reference/floating-point-ordering.md) | Compare the order of two floating point values |
+| [`isgreater`, `isgreaterequal`, `isless`, `islessequal`, `islessgreater`, `isunordered`](./reference/floating-point-ordering.md) | Compare the order of two floating-point values |
 | [`isinf`](./reference/isinf.md) | Determines whether a floating-point value is infinite |
 | [`isnan`, `_isnan`, `_isnanf`](./reference/isnan-isnan-isnanf.md) | Tests a floating-point value for NaN |
 | [`isnormal`](./reference/isnormal.md) | Tests whether a floating-point value is both finite and not subnormal |