Implement the ddx_fine HLSL Function

[farzonl]

• Implement ddx_fine clang builtin,
• Link ddx_fine clang builtin with hlsl_intrinsics.h
• Add sema checks for ddx_fine to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for ddx_fine to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/ddx_fine.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/ddx_fine-errors.hlsl
• Create the int_dx_ddx_fine intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_ddx_fine to 85 in DXIL.td
• Create the ddx_fine.ll and ddx_fine_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_ddx_fine intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the ddx_fine lowering and map it to int_spv_ddx_fine in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/ddx_fine.ll

DirectX

DXIL Opcode	DXIL OpName	Shader Model	Shader Stages
85	DerivFineX	6.0	('library', 'pixel', 'compute', 'amplification', 'mesh', 'node')

SPIR-V

OpDPdxFine:

Description:

Result is the partial derivative of P with respect to the window x
coordinate.Uses local differencing based on the value of P for the
current fragment and its immediate neighbor(s).

Result Type must be a scalar or vector of floating-point
type using the IEEE 754 encoding. The component width must
be 32 bits.

The type of P must be the same as Result Type. P is the value to
take the derivative of.

This instruction is only valid in the Fragment Execution
Model.

Capability:
DerivativeControl

Word Count	Opcode	Results	Operands
4	210	<id> Result Type	Result <id>	<id> P

Test Case(s)

Example 1

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

export float4 fn(float4 p1) {
    return ddx_fine(p1);
}

HLSL:

Computes a high precision partial derivative with respect to the screen-space x-coordinate.

Syntax

float ddx_fine(
  in float value
);

Parameters

value [in]

Type: float

The input value.

Return value

Type: float

The high precision partial derivative of value.

Remarks

The following overloaded versions are also available:

float2 ddx_fine(float2 value);
float3 ddx_fine(float3 value);
float4 ddx_fine(float4 value);

Minimum Shader Model

This function is supported in the following shader models.

Shader Model	Supported
Shader Model 5 and higher shader models	yes

 

This function is supported in the following types of shaders:

Vertex	Hull	Domain	Geometry	Pixel	Compute
				x

 

See also

Intrinsic Functions

Shader Model 5

[damyanp]

We should consider when implementing this microsoft/DirectXShaderCompiler#5744