fmax

Description

Floating-point maximum.

Signature

T fmax<T>(
    T x,
    T y)
    where T : __BuiltinFloatingPointType;

vector<T, N> fmax<T, int N>(
    vector<T, N> x,
    vector<T, N> y)
    where T : __BuiltinFloatingPointType;

Generic Parameters

T: __BuiltinFloatingPointType

N : int

Parameters

x : T

The first value to compare.

y : T

The second value to compare.

x : vector<T, N>

The first value to compare.

y : vector<T, N>

The second value to compare.

Return value

The larger of the two values, element-wise if vector typed.

Remarks

HLSL/DXIL returns the numeric operand for one NaN, and a NaN for two NaNs. SPIR-V uses the corresponding NMin/NMax operations; target floating-point modes may relax special-value behavior. GLSL uses native min/max, which need not follow those NaN rules. Zero-sign selection also follows the target operation.

Availability and Requirements

Defined for the following targets:

hlsl

Available in all stages.

glsl

Available in all stages.

cpp

Available in all stages.

cuda

Available in all stages.

metal

Available in all stages.

spirv

Available in all stages.

llvm

Available in all stages.