Implement the `isnan` HLSL Function

[farzonl]

• Implement isnan clang builtin,
• Link isnan clang builtin with hlsl_intrinsics.h
• Add sema checks for isnan to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for isnan to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/isnan.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/isnan-errors.hlsl
• Create the int_dx_isnan intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_isnan to 8 in DXIL.td
• Create the isnan.ll and isnan_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_isnan intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the isnan lowering and map it to int_spv_isnan in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/isnan.ll

DirectX

DXIL Opcode	DXIL OpName	Shader Model	Shader Stages
8	IsNaN	6.0	()

SPIR-V

OpIsNan:

Description:

Result is true if x is a NaN for the floating-point encoding used
by the type of x, otherwise result is false.

Result Type must be a scalar or vector of Boolean
type.

x must be a scalar or vector of floating-point type. It
must have the same number of components as Result Type.

Results are computed per component.

Word Count	Opcode	Results	Operands
4	156	<id> Result Type	Result <id>	<id> x

Test Case(s)

Example 1

//dxc isnan_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export bool4 fn(float4 p1) {
    return isnan(p1);
}

HLSL:

Determines if the specified value is NAN or QNAN.

ret isnan(x)

Parameters

Item	Description
x	[in] The specified value.

Return Value

Returns a value of the same size as the input, with a value set to True if the x parameter is NAN or QNAN. Otherwise, False.

Type Description

Name	Template Type	Component Type	Size
x	scalar, vector, or matrix	float	any
ret	scalar, vector, or matrix	bool	as input

Minimum Shader Model

This function is supported in the following shader models.

Shader Model	Supported
Shader Model 2 (DirectX HLSL) and higher shader models	yes
Shader Model 1 (DirectX HLSL)	yes (vs_1_1 only)

Requirements

Requirement	Value
Header	Corecrt_math.h

See also

Intrinsic Functions (DirectX HLSL)