Add support for (undocumented) HLSL 16-bit bit-cast ops (shader-slang#1528)

As of SM 6.2, the dxc compiler added support for a set of 16-bit bit-cast operations to mirror the `asuint`, `asfloat`, and `asint` operations that were provided for 32-bit scalar types. These operations are not publicly documented, so we didn't think to add them.

It should be noted that there was already a similar operation in HLSL, called `f32tof16`, that took as input a `float` and then packed a half-precision version of it into the low bits of a `uint`. The problem is that using that operation for `half`->`uint16_t` conversion required a round trip through a `float`, and downstream compilers seemingly can't optimize away that conversion.

This change adds the new operations along with a test that tries to make use of them to ensure the results are what is expected. There are enough cases to cover that I had to write the test in a way where each thread only writes out a subset of the required output.

There are two other changes here are that are not directly related to the main feature:

First, it seems like the `[__forceInlineEarly]` attribute on some of these overloads interacts poorly with generics, and results in an `IRVectorType` appearing at local scope in the output code. That is semantically reasonable given our IR model, but it would ideally be something that gets eliminated as a result of deduplication of types. For now I've introduced a slight hack to make types always get inlined into their use sites during emission, which should handle the case of locally-defined types. I'm not 100% happy with that solution, but it seemed better than introducing a bunch of unrelated fixes into this PR.

Second, the way that conversion operations were being declared for matrix types seems to have been incorrect: we had a single *explicit* initializer added to matrix types via an `extension` that allowed them to be initialized from other matrix types with the same size and *any* element type. In order to support implicit conversions of matrix types, I cribbed the code we were already using to introduce implicit conversion operations for vector types.

Author: Tim Foley

source/slang/core.meta.slang

@@ -520,13 +520,6 @@ for( int C = 2; C <= 4; ++C )
     }
     sb << ");\n";
 
-
-    // initialize from another matrix of the same size
-    //
-    // TODO(tfoley): See comment about how this overlaps
-    // with implicit conversion, in the `vector` case above
-    sb << "__generic<U> __init(matrix<U," << R << ", " << C << ">);\n";
-
     // initialize from a matrix of larger size
     for(int rr = R; rr <= 4; ++rr)
     for( int cc = C; cc <= 4; ++cc )
@@ -537,6 +530,33 @@ for( int C = 2; C <= 4; ++C )
 
     sb << "}\n";
 }
+
+for (int tt = 0; tt < kBaseTypeCount; ++tt)
+{
+    if(kBaseTypes[tt].tag == BaseType::Void) continue;
+    auto toType = kBaseTypes[tt].name;
+}}}}
+__generic<let R : int, let C : int> extension matrix<$(toType),R,C>
+{
+${{{{
+    for (int ff = 0; ff < kBaseTypeCount; ++ff)
+    {
+        if(kBaseTypes[ff].tag == BaseType::Void) continue;
+        if( tt == ff ) continue;
+
+        auto cost = getBaseTypeConversionCost(
+            kBaseTypes[tt],
+            kBaseTypes[ff]);
+        auto fromType = kBaseTypes[ff].name;
+}}}}
+    __implicit_conversion($(cost))
+    __init(matrix<$(fromType),R,C> value);
+${{{{
+    }
+}}}}
+}
+${{{{
+}
 }}}}
 
 

source/slang/hlsl.meta.slang

@@ -990,6 +990,71 @@ __generic<let N : int, let M : int>
 matrix<uint,N,M> asuint(matrix<uint,N,M> x)
 { return x; }
 
+
+// 16-bit bitcast ops (HLSL SM 6.2)
+//
+// TODO: We need to map these to GLSL/SPIR-V
+// operations that don't require an intermediate
+// conversion to fp32.
+
+// Identity cases:
+
+[__unsafeForceInlineEarly] float16_t asfloat16(float16_t value) { return value; }
+[__unsafeForceInlineEarly] vector<float16_t,N> asfloat16<let N : int>(vector<float16_t,N> value) { return value; }
+[__unsafeForceInlineEarly] matrix<float16_t,R,C> asfloat16<let R : int, let C : int>(matrix<float16_t,R,C> value) { return value; }
+
+[__unsafeForceInlineEarly] int16_t asint16(int16_t value) { return value; }
+[__unsafeForceInlineEarly] vector<int16_t,N> asint16<let N : int>(vector<int16_t,N> value) { return value; }
+[__unsafeForceInlineEarly] matrix<int16_t,R,C> asint16<let R : int, let C : int>(matrix<int16_t,R,C> value) { return value; }
+
+[__unsafeForceInlineEarly] uint16_t asuint16(uint16_t value) { return value; }
+[__unsafeForceInlineEarly] vector<uint16_t,N> asuint16<let N : int>(vector<uint16_t,N> value) { return value; }
+[__unsafeForceInlineEarly] matrix<uint16_t,R,C> asuint16<let R : int, let C : int>(matrix<uint16_t,R,C> value) { return value; }
+
+// Signed<->unsigned cases:
+
+[__unsafeForceInlineEarly] int16_t asint16(uint16_t value) { return value; }
+[__unsafeForceInlineEarly] vector<int16_t,N> asint16<let N : int>(vector<uint16_t,N> value) { return value; }
+[__unsafeForceInlineEarly] matrix<int16_t,R,C> asint16<let R : int, let C : int>(matrix<uint16_t,R,C> value) { return value; }
+
+[__unsafeForceInlineEarly] uint16_t asuint16(int16_t value) { return value; }
+[__unsafeForceInlineEarly] vector<uint16_t,N> asuint16<let N : int>(vector<int16_t,N> value) { return value; }
+[__unsafeForceInlineEarly] matrix<uint16_t,R,C> asuint16<let R : int, let C : int>(matrix<int16_t,R,C> value) { return value; }
+
+// Float->unsigned cases:
+
+__target_intrinsic(hlsl)
+__target_intrinsic(glsl, "uint16_t(packHalf2x16(vec2($0, 0.0)))")
+uint16_t asuint16(float16_t value);
+
+vector<uint16_t,N> asuint16<let N : int>(vector<float16_t,N> value)
+{ VECTOR_MAP_UNARY(uint16_t, N, asuint16, value); }
+
+matrix<uint16_t,R,C> asuint16<let R : int, let C : int>(matrix<float16_t,R,C> value)
+{ MATRIX_MAP_UNARY(uint16_t, R, C, asuint16, value); }
+
+// Unsigned->float cases:
+
+__target_intrinsic(hlsl)
+__target_intrinsic(glsl, "float16_t(unpackHalf2x16($0).x)")
+float16_t asfloat16(uint16_t value);
+
+vector<float16_t,N> asfloat16<let N : int>(vector<uint16_t,N> value)
+{ VECTOR_MAP_UNARY(float16_t, N, asfloat16, value); }
+
+matrix<float16_t,R,C> asfloat16<let R : int, let C : int>(matrix<uint16_t,R,C> value)
+{ MATRIX_MAP_UNARY(float16_t, R, C, asfloat16, value); }
+
+// Float<->signed cases:
+
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] int16_t asint16(float16_t value) { return asuint16(value); }
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] vector<int16_t,N> asint16<let N : int>(vector<float16_t,N> value) { return asuint16(value); }
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] matrix<int16_t,R,C> asint16<let R : int, let C : int>(matrix<float16_t,R,C> value) { return asuint16(value); }
+
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] float16_t asfloat16(int16_t value) { return asfloat16(asuint16(value)); }
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] vector<float16_t,N> asfloat16<let N : int>(vector<int16_t,N> value) { return asfloat16(asuint16(value)); }
+__target_intrinsic(hlsl) [__unsafeForceInlineEarly] matrix<float16_t,R,C> asfloat16<let R : int, let C : int>(matrix<int16_t,R,C> value) { return asfloat16(asuint16(value)); }
+
 // Inverse tangent (HLSL SM 1.0)
 __generic<T : __BuiltinFloatingPointType>
 __target_intrinsic(hlsl)

source/slang/slang-emit-c-like.cpp

@@ -974,6 +974,18 @@ bool CLikeSourceEmitter::shouldFoldInstIntoUseSites(IRInst* inst)
     // for temporary variables.
     auto type = inst->getDataType();
 
+    // We treat instructions that yield a type as things we should *always* fold.
+    //
+    // TODO: In general, at the point where we emit code we do not expect to
+    // find types being constructed locally (inside function bodies), but this
+    // can end up happening because of interaction between different features.
+    // Notably, if a generic function gets force-inlined early in codegen,
+    // then any types it constructs will be inlined into the body of the caller
+    // by default.
+    //
+    if(as<IRType>(inst) || as<IRTypeKind>(type))
+        return true;
+
     // Unwrap any layers of array-ness from the type, so that
     // we can look at the underlying data type, in case we
     // should *never* expose a value of that type

tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang

@@ -0,0 +1,103 @@
+// bit-cast-16-bit.slang
+
+//TEST(compute):COMPARE_COMPUTE_EX:-slang -compute -dx12 -use-dxil -profile sm_6_2
+//TEST(compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute
+
+//TEST_INPUT:ubuffer(data=[0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16], stride=4):name inputBuffer
+RWStructuredBuffer<int> inputBuffer;
+
+int16_t readI(inout int index) { return inputBuffer[(index++) & 0xF]; }
+uint16_t readU(inout int index) { return inputBuffer[(index++) & 0xF]; }
+float16_t readF(inout int index) { return float(inputBuffer[(index++) & 0xF]); }
+
+//TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0 0 0 0 0], stride=4):out,name outputBuffer
+RWStructuredBuffer<int> outputBuffer;
+
+void write(int initial, inout int index, uint value)
+{
+    let tmp = index++;
+    if((tmp & 3) == initial)
+    {
+        outputBuffer[tmp & 0xF] = value;
+    }
+}
+
+void write(int initial, inout int index, uint16_t value)
+{
+    write(initial, index, uint(value));
+}
+
+void write(int initial, inout int index, int16_t value)
+{
+    write(initial, index, uint(int(value)));
+}
+
+void write(int initial, inout int index, float16_t value)
+{
+    write(initial, index, asuint(float(value)));
+}
+
+
+void test(int initial)
+{
+    int input = initial;
+    int output = 0;
+
+    // Scalar
+    {
+        let i = readI(input);
+        let u = readU(input);
+        let f = readF(input);
+
+        // int->float
+        let a = asfloat16(i);
+        write(initial, output, a);
+
+        // float->uint
+        let b = asuint16(f);
+        write(initial, output, b);
+
+        // uint->int
+        let c = asint16(u);
+        write(initial, output, c);
+
+        // float->float
+        let d = asfloat16(f);
+        write(initial, output, d);
+    }
+
+    // Vector
+    {
+        let i = int16_t2(readI(input), readI(input));
+        let u = uint16_t2(readU(input), readU(input));
+        let f = float16_t2(readF(input), readF(input));
+
+        // uint->float
+        let a = asfloat16(u);
+        write(initial, output, a.x);
+        write(initial, output, a.y);
+
+        // float->int
+        let b = asint16(f);
+        write(initial, output, b.x);
+        write(initial, output, b.y);
+
+        // int->uint
+        let c = asuint16(i);
+        write(initial, output, c.x);
+        write(initial, output, c.y);
+
+        // int->int
+        let d = asint16(i);
+        write(initial, output, d.x);
+        write(initial, output, d.y);
+    }
+
+}
+
+
+[numthreads(4, 1, 1)]
+void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
+{
+    test(dispatchThreadID.x);
+}

tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang.expected.txt

@@ -0,0 +1,12 @@
+0
+4200
+3
+40A00000
+34A00000
+34E00000
+4880
+4980
+3
+5
+5
+7