Implement the `dot2add` HLSL Function

[farzonl]

• Implement dot2add clang builtin,
• Link dot2add clang builtin with hlsl_intrinsics.h
• Add sema checks for dot2add to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for dot2add to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/dot2add.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/dot2add-errors.hlsl
• Create the int_dx_dot2add intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_dot2add to 162 in DXIL.td
• Create the dot2add.ll and dot2add_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_dot2add intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the dot2add lowering and map it to int_spv_dot2add in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/dot2add.ll

DirectX

DXIL Opcode	DXIL OpName	Shader Model	Shader Stages
162	Dot2AddHalf	6.4	()

SPIR-V

OpDot:

Description:

Dot product of Vector 1 and Vector 2.

Result Type must be a floating-point type scalar.

Vector 1 and Vector 2 must be vectors of the same type, and their
component type must be Result Type.

Word Count	Opcode	Results	Operands
5	148	<id> Result Type	Result <id>	<id> Vector 1	<id> Vector 2

Test Case(s)

Example 1

//dxc dot2add_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float fn(half2 p1, half2 p2, float p3) {
    return dot2add(p1, p2, p3);
}

HLSL:

Syntax

float dot2add(float16_t<2> a, float16_t<2> b, float c);

Type Description

Name	Template Type	Component Type	Size
ret	scalar	float	1
a	vector	half	2
b	vector	half	2
c	scalar	float	1

Minimum Shader Model

This function is supported in the following shader models.

Shader Model	Supported
Shader Model 6.4 and higher shader models	yes

Shader Stages

See also

• Intrinsic Functions (DirectX HLSL)
• See Single Precision Floating Point 2-Element Dot-Product and Accumulate

[sumitsays]

Will start working on this.