Implement the `WaveActiveMin` HLSL Function

[farzonl]

• Implement WaveActiveMin clang builtin,
• Link WaveActiveMin clang builtin with hlsl_intrinsics.h
• Add sema checks for WaveActiveMin to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for WaveActiveMin to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/WaveActiveMin.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/WaveActiveMin-errors.hlsl
• Create the int_dx_WaveActiveMin intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_WaveActiveMin to 119 in DXIL.td
• Create the WaveActiveMin.ll and WaveActiveMin_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_WaveActiveMin intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the WaveActiveMin lowering and map it to int_spv_WaveActiveMin in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveMin.ll

DirectX

DXIL Opcode	DXIL OpName	Shader Model	Shader Stages
119	WaveActiveOp	6.0	('library', 'compute', 'amplification', 'mesh', 'pixel', 'vertex', 'hull', 'domain', 'geometry', 'raygeneration', 'intersection', 'anyhit', 'closesthit', 'miss', 'callable', 'node')

SPIR-V

OpGroupNonUniformFMin:

Description:

A floating point minimum group operation of all
Value operands contributed by active invocations in the
group.

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

Execution is a Scope that identifies the group of
invocations affected by this command. It must be Subgroup.

The identity I for Operation is +INF. If Operation is
ClusteredReduce, ClusterSize must be present.

The type of Value must be the same as Result Type. The method used
to perform the group operation on the contributed Value(s) from active
invocations is implementation defined. From the set of Value(s)
provided by active invocations within a subgroup, if for any two
Values one of them is a NaN, the other is chosen. If all Value(s)
that are used by the current invocation are NaN, then the result is an
undefined value.

ClusterSize is the size of cluster to use. ClusterSize must be a
scalar of integer type, whose Signedness operand is 0.
ClusterSize must come from a constant
instruction. Behavior is undefined unless
ClusterSize is at least 1 and a power of 2. If ClusterSize is
greater than the size of the group, executing this instruction
results in undefined behavior.

Capability:
GroupNonUniformArithmetic, GroupNonUniformClustered,
GroupNonUniformPartitionedNV

Missing before version 1.3.

Word Count	Opcode	Results	Operands
6 + variable	355	<id> Result Type	Result <id>	Scope <id> Execution	Group Operation Operation	<id> Value	Optional <id> ClusterSize

Test Case(s)

Example 1

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

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

Example 2

//dxc WaveActiveMin_1_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export uint4 fn(uint4 p1) {
    return WaveActiveMin(p1);
}

Example 3

//dxc WaveActiveMin_2_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int4 fn(int4 p1) {
    return WaveActiveMin(p1);
}

HLSL:

Returns the minimum value of the expression across all active lanes in the current wave replicates it back to all active lanes.

Syntax

<type> WaveActiveMin(
   <type> expr
);

Parameters

expr

The expression to evaluate.

Return value

The minimum value.

Remarks

The order of operations is undefined.

This function is supported from shader model 6.0 in all shader stages.

 

Examples

 float3 minPos = WaveActiveMin( myPoint.position );
    BoundingBox.min = min( minPos, BoundingBox.min );

See also

Overview of Shader Model 6

Shader Model 6

[KungFuDonkey]

I will look into this

[damyanp]

Hi! Thank you for offering to look into this. As we add new HLSL functions like this (aka intrinsics), we want to add offload-test-suites for them first (or at least simultaneously). We've just completed an exercise where we went back through all the functions we implemented previously and added new offload-test-suite tests for them.

If you look through completed PRs under https://github.com/llvm/offload-test-suite you can see some examples. This one looks like it might be particularly relevant to this issue: llvm/offload-test-suite#429

[KungFuDonkey]

Thanks, I just went through the checklist, and followed mostly what WaveActiveMax was doing in llvm for the tests. I guess it does not hurt to also add the change there.

Is the offload-test-suite an external project? I cannot find it inside the llvm-project

[KungFuDonkey]

Oh I see, its an external one that is added to the LLVM_EXTERNAL_PROJECT variable