Implement the `mul` HLSL Function

[farzonl]

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

DirectX

There were no DXIL opcodes found for mul.

SPIR-V

OpFMul:

Description:

Floating-point multiplication of Operand 1 and Operand 2.

Result Type must be a scalar or vector of floating-point
type.

The types of Operand 1 and Operand 2 both must be the same as
Result Type.

Results are computed per component.

Word Count	Opcode	Results	Operands
5	133	<id> Result Type	Result <id>	<id> Operand 1	<id> Operand 2

Test Case(s)

Example 1

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

export float fn(float p1, float p2) {
    return mul(p1, p2);
}

Example 2

//dxc mul_1_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint fn(uint p1, uint p2) {
    return mul(p1, p2);
}

Example 3

//dxc mul_2_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int fn(int p1, int p2) {
    return mul(p1, p2);
}

Example 4

//dxc mul_3_test.hlsl -T lib_6_8  -enable-16bit-types -spirv -fspv-target-env=universal1.5 -fcgl -O0

export float4 fn(float4 p1, float p2) {
    return mul(p1, p2);
}

Example 5

//dxc mul_4_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4 fn(float p1, float4 p2) {
    return mul(p1, p2);
}

Example 6

//dxc mul_5_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4 fn(uint p1, uint4 p2) {
    return mul(p1, p2);
}

Example 7

//dxc mul_6_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4 fn(int p1, int4 p2) {
    return mul(p1, p2);
}

Example 8

//dxc mul_7_test.hlsl -T lib_6_8  -enable-16bit-types -spirv -fspv-target-env=universal1.5 -fcgl -O0

export float4 fn(float4x4 p1, float4 p2) {
    return mul(p1, p2);
}

Example 9

//dxc mul_8_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4x4 fn(float p1, float4x4 p2) {
    return mul(p1, p2);
}

Example 10

//dxc mul_9_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4x4 fn(uint p1, uint4x4 p2) {
    return mul(p1, p2);
}

Example 11

//dxc mul_10_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4x4 fn(int p1, int4x4 p2) {
    return mul(p1, p2);
}

Example 12

//dxc mul_12_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4 fn(float4 p1, float p2) {
    return mul(p1, p2);
}

Example 13

//dxc mul_13_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4 fn(uint4 p1, uint p2) {
    return mul(p1, p2);
}

Example 14

//dxc mul_14_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4 fn(int4 p1, int p2) {
    return mul(p1, p2);
}

Example 15

//dxc mul_16_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float fn(float4 p1, float4 p2) {
    return mul(p1, p2);
}

Example 16

//dxc mul_17_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint fn(uint4 p1, uint4 p2) {
    return mul(p1, p2);
}

Example 17

//dxc mul_18_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int fn(int4 p1, int4 p2) {
    return mul(p1, p2);
}

Example 18

//dxc mul_20_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4 fn(float4 p1, float4x4 p2) {
    return mul(p1, p2);
}

Example 19

//dxc mul_21_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4 fn(uint4 p1, uint4x4 p2) {
    return mul(p1, p2);
}

Example 20

//dxc mul_22_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4 fn(int4 p1, int4x4 p2) {
    return mul(p1, p2);
}

Example 21

//dxc mul_24_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4x4 fn(float4x4 p1, float p2) {
    return mul(p1, p2);
}

Example 22

//dxc mul_25_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4x4 fn(uint4x4 p1, uint p2) {
    return mul(p1, p2);
}

Example 23

//dxc mul_26_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4x4 fn(int4x4 p1, int p2) {
    return mul(p1, p2);
}

Example 24

//dxc mul_28_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4 fn(float4x4 p1, float4 p2) {
    return mul(p1, p2);
}

Example 25

//dxc mul_29_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4 fn(uint4x4 p1, uint4 p2) {
    return mul(p1, p2);
}

Example 26

//dxc mul_30_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4 fn(int4x4 p1, int4 p2) {
    return mul(p1, p2);
}

Example 27

//dxc mul_32_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export float4x4 fn(float4x4 p1, float4x4 p2) {
    return mul(p1, p2);
}

Example 28

//dxc mul_33_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4x4 fn(uint4x4 p1, uint4x4 p2) {
    return mul(p1, p2);
}

Example 29

//dxc mul_34_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4x4 fn(int4x4 p1, int4x4 p2) {
    return mul(p1, p2);
}

HLSL:

Multiplies x and y using matrix math. The inner dimension x-columns and y-rows must be equal.

ret mul(x, y)

Parameters

Item	Description
x	[in] The x input value. If x is a vector, it treated as a row vector.
y	[in] The y input value. If y is a vector, it treated as a column vector.

Return Value

The result of x times y. The result has the dimension x-rows x y-columns.

Type Description

There are 9 overloaded versions of this function; the overloaded versions handle the different cases for the types and sizes of the input arguments.

Version	Name	Purpose	Template Type	Component Type	Size
1
	x	in	scalar	float, int	1
	y	in	scalar	same as input x	1
	ret	out	scalar	same as input x	1
2
	x	in	scalar	float, int	1
	y	in	vector	float, int	any
	ret	out	vector	float, int	same dimension(s) as input y
3
	x	in	scalar	float, int	1
	y	in	matrix	float, int	any
	ret	out	matrix	same as input y	same dimension(s) as input y
4
	x	in	vector	float, int	any
	y	in	scalar	float, int	1
	ret	out	vector	float, int	same dimension(s) as input x
5
	x	in	vector	float, int	any
	y	in	vector	float, int	same dimension(s) as input x
	ret	out	scalar	float, int	1
6
	x	in	vector	float, int	any
	y	in	matrix	float, int	rows = same dimension(s) as input x, columns = any
	ret	out	vector	float, int	same dimension(s) as input y columns
7
	x	in	matrix	float, int	any
	y	in	scalar	float, int	1
	ret	out	matrix	float, int	same dimension(s) as input x
8
	x	in	matrix	float, int	any
	y	in	vector	float, int	number of columns in input x
	ret	out	vector	float, int	number of rows in input x
9
	x	in	matrix	float, int	any
	y	in	matrix	float, int	rows = number of columns in input x
	ret	out	matrix	float, int	rows = number of rows in input x, columns = number of columns in input y

Minimum Shader Model

This function is supported in the following shader models.

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

See also

Intrinsic Functions (DirectX HLSL)

[pow2clk]

Depends on matrix support. #109839 at least. Some complication is that the existing matrix extension performs matrix multiply with the * operator while HLSL uses that for elementwise multiplication