Port core scan shaders and some other fixes. Missing descirptors for full scan port

Author: kpentaris

include/nbl/builtin/hlsl/scan/direct.hlsl

@@ -0,0 +1,50 @@
+#ifndef _NBL_HLSL_WORKGROUP_SIZE_
+#define _NBL_HLSL_WORKGROUP_SIZE_ 256
+#endif
+
+#include <nbl/builtin/hlsl/scan/descriptors.hlsl>
+#include <nbl/builtin/hlsl/scan/virtual_workgroup.hlsl>
+#include <nbl/builtin/hlsl/scan/default_scheduler.hlsl>
+
+namespace nbl
+{
+namespace hlsl
+{
+namespace scan
+{
+#ifndef _NBL_HLSL_SCAN_PUSH_CONSTANTS_DEFINED_
+	cbuffer PC // REVIEW: register and packoffset selection
+	{
+		Parameters_t scanParams;
+		DefaultSchedulerParameters_t schedulerParams;
+	};
+#define _NBL_HLSL_SCAN_PUSH_CONSTANTS_DEFINED_
+#endif
+
+#ifndef _NBL_HLSL_SCAN_GET_PARAMETERS_DEFINED_
+Parameters_t getParameters()
+{
+	return pc.scanParams;
+}
+#define _NBL_HLSL_SCAN_GET_PARAMETERS_DEFINED_
+#endif
+
+#ifndef _NBL_HLSL_SCAN_GET_SCHEDULER_PARAMETERS_DEFINED_
+DefaultSchedulerParameters_t getSchedulerParameters()
+{
+	return pc.schedulerParams;
+}
+#define _NBL_HLSL_SCAN_GET_SCHEDULER_PARAMETERS_DEFINED_
+#endif
+}
+}
+}
+
+#ifndef _NBL_HLSL_MAIN_DEFINED_
+[numthreads(_NBL_HLSL_WORKGROUP_SIZE_, 1, 1)]
+void CSMain()
+{
+	nbl::hlsl::scan::main();
+}
+#define _NBL_HLSL_MAIN_DEFINED_
+#endif

include/nbl/builtin/hlsl/scan/indirect.hlsl

@@ -0,0 +1,48 @@
+#ifndef _NBL_HLSL_WORKGROUP_SIZE_
+#define _NBL_HLSL_WORKGROUP_SIZE_ 256
+#define _NBL_HLSL_WORKGROUP_SIZE_LOG2_ 8
+#endif
+
+#include <nbl/builtin/hlsl/scan/descriptors.hlsl>
+#include <nbl/builtin/hlsl/scan/virtual_workgroup.hlsl>
+#include <nbl/builtin/hlsl/scan/default_scheduler.hlsl>
+
+namespace nbl
+{
+namespace hlsl
+{
+namespace scan
+{
+#ifndef _NBL_HLSL_SCAN_GET_PARAMETERS_DEFINED_
+Parameters_t scanParams;
+Parameters_t getParameters()
+{
+	return scanParams;
+}
+#define _NBL_HLSL_SCAN_GET_PARAMETERS_DEFINED_
+#endif
+
+uint getIndirectElementCount();
+
+#ifndef _NBL_HLSL_SCAN_GET_SCHEDULER_PARAMETERS_DEFINED_
+DefaultSchedulerParameters_t schedulerParams;
+DefaultSchedulerParameters_t getSchedulerParameters()
+{
+	scheduler::computeParameters(getIndirectElementCount(),scanParams,schedulerParams);
+	return schedulerParams;
+}
+#define _NBL_HLSL_SCAN_GET_SCHEDULER_PARAMETERS_DEFINED_
+#endif
+}
+}
+}
+
+#ifndef _NBL_HLSL_MAIN_DEFINED_
+[numthreads(_NBL_HLSL_WORKGROUP_SIZE_, 1, 1)]
+void CSMain()
+{
+	if (bool(nbl::hlsl::scan::getIndirectElementCount()))
+		nbl::hlsl::scan::main();
+}
+#define _NBL_HLSL_MAIN_DEFINED_
+#endif

include/nbl/builtin/hlsl/scan/parameters_struct.hlsl

@@ -13,9 +13,10 @@ namespace hlsl
 {
 namespace scan
 {
+	// REVIEW: Putting topLevel second allows better alignment for packing of constant variables, assuming lastElement has length 4. (https://learn.microsoft.com/en-us/windows/win32/direct3dhlsl/dx-graphics-hlsl-packing-rules)
 	struct Parameters_t {
-		uint topLevel;
 		uint lastElement[NBL_BUILTIN_MAX_SCAN_LEVELS/2+1];
+		uint topLevel;
 		uint temporaryStorageOffset[NBL_BUILTIN_MAX_SCAN_LEVELS/2];
 	}
 }

include/nbl/builtin/hlsl/scan/virtual_workgroup.hlsl

@@ -30,7 +30,7 @@ namespace scan
 
 		const bool inRange = levelInvocationIndex <= params.lastElement[pseudoLevel];
 
-		Storage_t data = binop::identity();
+		Storage_t data = binop.identity();
 		if(inRange)
 		{
 			getData(data, levelInvocationIndex, localWorkgroupIndex, treeLevel, pseudoLevel);

include/nbl/builtin/hlsl/subgroup/arithmetic_portability_impl.hlsl

@@ -318,7 +318,7 @@ struct inclusive_scan : scan_base
 			// additionally, the first half invocations initialize the padding slots
 			// with identity values
 			if (scan_base::pseudoSubgroupInvocation<scan_base::HalfSubgroupSize)
-				scratchAccessor.set(lastLoadOffset,op::identity());
+				scratchAccessor.set(lastLoadOffset,op.identity());
 		}
 		nbl::hlsl::subgroup::Barrier();
 		nbl::hlsl::subgroup::MemoryBarrierShared();