[CUDA] Fix incorrect `kIROp_RaytracingAccelerationStructureType` emitting logic (#8168)

Fixes: #8167

Current emitting logic does not work, this has been corrected.
The provided test ensures our CUDA code is valid by compiling PTX from
it.

`m_writer->emit("OptixTraversableHandle");` should be `out <<` since
`out` adds to type-name-cache; otherwise using a type twice will produce
bad type-names (since we filled type-name cache with "" instead of
"typeName")

2 files changed, 2 insertions, 17 deletions

diff --git a/source/slang/slang-emit-cuda.cpp b/source/slang/slang-emit-cuda.cpp
index e27fd25aa..a5b90740d 100644
--- a/source/slang/slang-emit-cuda.cpp
+++ b/source/slang/slang-emit-cuda.cpp
@@ -214,6 +214,7 @@ SlangResult CUDASourceEmitter::calcTypeName(IRType* type, CodeGenTarget target,
             out << "TensorView";
             return SLANG_OK;
         }
+    case kIROp_RaytracingAccelerationStructureType:
     case kIROp_HitObjectType:
         {
             out << "OptixTraversableHandle";
@@ -254,22 +255,6 @@ SlangResult CUDASourceEmitter::calcTypeName(IRType* type, CodeGenTarget target,
         }
     }
 
-    if (auto untypedBufferType = as<IRUntypedBufferResourceType>(type))
-    {
-        switch (untypedBufferType->getOp())
-        {
-        case kIROp_RaytracingAccelerationStructureType:
-            {
-                m_writer->emit("OptixTraversableHandle");
-                return SLANG_OK;
-                break;
-            }
-
-        default:
-            break;
-        }
-    }
-
     return Super::calcTypeName(type, target, out);
 }
 
diff --git a/source/slang/slang-intrinsic-expand.cpp b/source/slang/slang-intrinsic-expand.cpp
index 56ff2a108..28023e9e2 100644
--- a/source/slang/slang-intrinsic-expand.cpp
+++ b/source/slang/slang-intrinsic-expand.cpp
@@ -446,7 +446,7 @@ const char* IntrinsicExpandContext::_emitSpecial(const char* cursor)
             // there is a format conversion required between the type associated by the resource and
             // the backing ImageFormat. Currently this is only implemented on CUDA, where there are
             // specialized versions of the RWTexture writes that will do a format conversion.
-            if (m_emitter->getTarget() == CodeGenTarget::CUDASource)
+            if (isCUDATarget(m_emitter->getTargetReq()))
             {
                 IRInst* resourceInst = m_callInst->getArg(0);