Implement the `ddx` HLSL Function

[farzonl]

• Implement ddx clang builtin,
• Link ddx clang builtin with hlsl_intrinsics.h
• Add sema checks for ddx to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for ddx to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/ddx.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/ddx-errors.hlsl
• Create the int_dx_ddx intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_ddx to 83 in DXIL.td
• Create the ddx.ll and ddx_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_ddx intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the ddx lowering and map it to int_spv_ddx in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/ddx.ll

DirectX

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

SPIR-V

OpDPdx:

Description:

Same result as either OpDPdxFine or
OpDPdxCoarse on P. Selection of which one is
based on external factors.

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:
Shader

Word Count	Opcode	Results	Operands
4	207	<id> Result Type	Result <id>	<id> P

Test Case(s)

Example 1

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

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

HLSL:

Returns the partial derivative of the specified value with respect to the screen-space x-coordinate.

ret ddx(x)

This function computes the partial derivative with respect to the screen-space x-coordinate. To compute the partial derivative with respect to the screen-space y-coordinate, use the ddy function.

This function is only supported in pixel shaders.

Parameters

Item	Description
x	[in] The specified value.

Return Value

The partial derivative of the x parameter.

Type Description

Name	Template Type	Component Type	Size
x	scalar, vector, or matrix	float	any
ret	same as input x	float	same dimension(s) as input x

Minimum Shader Model

This function is supported in the following shader models.

Shader Model	Supported
Shader Model 5 and higher shader models	yes
Shader Model 4	yes
Shader Model 3 (DirectX HLSL)	yes
Shader Model 2 (DirectX HLSL)	yes in ps_2_x; unsupported in ps_2_0.
Shader Model 1 (DirectX HLSL)	no

This function is supported in the following types of shaders:

Vertex	Hull	Domain	Geometry	Pixel	Compute
				x

See also

Intrinsic Functions (DirectX HLSL)

[damyanp]

We should consider when implementing this microsoft/DirectXShaderCompiler#5744