Add support for (undocumented) HLSL 16-bit bit-cast ops (#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.

2 files changed, 115 insertions, 0 deletions

diff --git a/tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang b/tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang
new file mode 100644
index 000000000..66f015172
--- /dev/null
+++ b/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);
+}
diff --git a/tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang.expected.txt b/tests/hlsl-intrinsic/bit-cast/bit-cast-16-bit.slang.expected.txt
new file mode 100644
index 000000000..5d8824457
--- /dev/null
+++ b/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