diff --git a/WickedEngine/wiEmittedParticle.cpp b/WickedEngine/wiEmittedParticle.cpp
index f195befda..cf6352467 100644
--- a/WickedEngine/wiEmittedParticle.cpp
+++ b/WickedEngine/wiEmittedParticle.cpp
@@ -313,12 +313,8 @@ namespace wi
 		std::swap(aliveList[0], aliveList[1]);
 
 		// Read back statistics (with GPU delay):
-		if (statisticsReadBackIndex > arraysize(statisticsReadbackBuffer))
-		{
-			const uint32_t oldest_stat_index = (statisticsReadBackIndex + 1) % arraysize(statisticsReadbackBuffer);
-			memcpy(&statistics, statisticsReadbackBuffer[oldest_stat_index].mapped_data, sizeof(statistics));
-		}
-		statisticsReadBackIndex++;
+		const uint32_t oldest_stat_index = wi::graphics::GetDevice()->GetBufferIndex();
+		memcpy(&statistics, statisticsReadbackBuffer[oldest_stat_index].mapped_data, sizeof(statistics));
 	}
 	void EmittedParticleSystem::Burst(int num)
 	{
@@ -679,7 +675,8 @@ namespace wi
 		}
 
 		// Statistics is copied to readback:
-		device->CopyResource(&statisticsReadbackBuffer[(statisticsReadBackIndex - 1) % arraysize(statisticsReadbackBuffer)], &counterBuffer, cmd);
+		const uint32_t oldest_stat_index = wi::graphics::GetDevice()->GetBufferIndex();
+		device->CopyResource(&statisticsReadbackBuffer[oldest_stat_index], &counterBuffer, cmd);
 
 		{
 			const GPUBarrier barriers[] = {
diff --git a/WickedEngine/wiEmittedParticle.h b/WickedEngine/wiEmittedParticle.h
index 4519d695f..385a2d9b3 100644
--- a/WickedEngine/wiEmittedParticle.h
+++ b/WickedEngine/wiEmittedParticle.h
@@ -31,7 +31,7 @@ namespace wi
 		};
 
 		ParticleCounters statistics = {};
-		wi::graphics::GPUBuffer statisticsReadbackBuffer[wi::graphics::GraphicsDevice::GetBufferCount() + 1];
+		wi::graphics::GPUBuffer statisticsReadbackBuffer[wi::graphics::GraphicsDevice::GetBufferCount()];
 
 		wi::graphics::GPUBuffer particleBuffer;
 		wi::graphics::GPUBuffer aliveList[2];
@@ -129,7 +129,6 @@ namespace wi
 
 		// Non-serialized attributes:
 		XMFLOAT3 center;
-		uint32_t statisticsReadBackIndex = 0;
 		uint32_t layerMask = ~0u;
 
 		inline bool IsDebug() const { return _flags & FLAG_DEBUG; }
diff --git a/WickedEngine/wiGraphicsDevice.h b/WickedEngine/wiGraphicsDevice.h
index 146188bc2..00f998112 100644
--- a/WickedEngine/wiGraphicsDevice.h
+++ b/WickedEngine/wiGraphicsDevice.h
@@ -49,7 +49,7 @@ namespace wi::graphics
 	class GraphicsDevice
 	{
 	protected:
-		static const uint32_t BUFFERCOUNT = 2;
+		static constexpr uint32_t BUFFERCOUNT = 2;
 		uint64_t FRAMECOUNT = 0;
 		ValidationMode validationMode = ValidationMode::Disabled;
 		GraphicsDeviceCapability capabilities = GraphicsDeviceCapability::NONE;
@@ -57,8 +57,8 @@ namespace wi::graphics
 		size_t TOPLEVEL_ACCELERATION_STRUCTURE_INSTANCE_SIZE = 0;
 		uint32_t VARIABLE_RATE_SHADING_TILE_SIZE = 0;
 		uint64_t TIMESTAMP_FREQUENCY = 0;
-		uint32_t vendorId;
-		uint32_t deviceId;
+		uint32_t vendorId = 0;
+		uint32_t deviceId = 0;
 		std::string adapterName;
 		std::string driverDescription;
 		AdapterType adapterType = AdapterType::Other;
@@ -119,7 +119,7 @@ namespace wi::graphics
 		// Returns the buffer count, which is the array size of buffered resources used by both the CPU and GPU
 		static constexpr uint32_t GetBufferCount() { return BUFFERCOUNT; }
 		// Returns the current buffer index, which is in range [0, GetBufferCount() - 1]
-		constexpr uint32_t GetBufferIndex() const { return GetFrameCount() % BUFFERCOUNT; }
+		constexpr uint32_t GetBufferIndex() const { return GetFrameCount() % GetBufferCount(); }
 
 		// Returns whether the graphics debug layer is enabled. It can be enabled when creating the device.
 		constexpr bool IsDebugDevice() const { return validationMode != ValidationMode::Disabled; }
diff --git a/WickedEngine/wiGraphicsDevice_DX12.cpp b/WickedEngine/wiGraphicsDevice_DX12.cpp
index 5a276d51c..ebf64b260 100644
--- a/WickedEngine/wiGraphicsDevice_DX12.cpp
+++ b/WickedEngine/wiGraphicsDevice_DX12.cpp
@@ -5072,24 +5072,22 @@ using namespace dx12_internal;
 		// From here, we begin a new frame, this affects GetBufferIndex()!
 		FRAMECOUNT++;
 
-		// Begin next frame:
+		// Initiate stalling CPU when GPU is not yet finished with next frame:
+		const uint32_t bufferindex = GetBufferIndex();
+		for (int queue = 0; queue < QUEUE_COUNT; ++queue)
 		{
-			// Initiate stalling CPU when GPU is not yet finished with next frame:
-			for (int queue = 0; queue < QUEUE_COUNT; ++queue)
+			if (FRAMECOUNT >= BUFFERCOUNT && frame_fence[bufferindex][queue]->GetCompletedValue() < 1)
 			{
-				if (FRAMECOUNT >= BUFFERCOUNT && frame_fence[GetBufferIndex()][queue]->GetCompletedValue() < 1)
-				{
-					// NULL event handle will simply wait immediately:
-					//	https://docs.microsoft.com/en-us/windows/win32/api/d3d12/nf-d3d12-id3d12fence-seteventoncompletion#remarks
-					hr = frame_fence[GetBufferIndex()][queue]->SetEventOnCompletion(1, NULL);
-					assert(SUCCEEDED(hr));
-				}
-				hr = frame_fence[GetBufferIndex()][queue]->Signal(0);
+				// NULL event handle will simply wait immediately:
+				//	https://docs.microsoft.com/en-us/windows/win32/api/d3d12/nf-d3d12-id3d12fence-seteventoncompletion#remarks
+				hr = frame_fence[bufferindex][queue]->SetEventOnCompletion(1, NULL);
+				assert(SUCCEEDED(hr));
 			}
-			assert(SUCCEEDED(hr));
-
-			allocationhandler->Update(FRAMECOUNT, BUFFERCOUNT);
+			hr = frame_fence[bufferindex][queue]->Signal(0);
 		}
+		assert(SUCCEEDED(hr));
+
+		allocationhandler->Update(FRAMECOUNT, BUFFERCOUNT);
 	}
 
 	void GraphicsDevice_DX12::OnDeviceRemoved()
diff --git a/WickedEngine/wiProfiler.cpp b/WickedEngine/wiProfiler.cpp
index 1d85187ca..487a06036 100644
--- a/WickedEngine/wiProfiler.cpp
+++ b/WickedEngine/wiProfiler.cpp
@@ -32,9 +32,9 @@ namespace wi::profiler
 	range_id cpu_frame;
 	range_id gpu_frame;
 	GPUQueryHeap queryHeap;
-	GPUBuffer queryResultBuffer[GraphicsDevice::GetBufferCount() + 1];
+	GPUBuffer queryResultBuffer[GraphicsDevice::GetBufferCount()];
 	std::atomic<uint32_t> nextQuery{ 0 };
-	int queryheap_idx = 0;
+	uint32_t queryheap_idx = 0;
 	bool drawn_this_frame = false;
 	wi::Color background_color = wi::Color(20, 20, 20, 230);
 	wi::Color text_color = wi::Color::White();
@@ -58,6 +58,9 @@ namespace wi::profiler
 		int gpuBegin[arraysize(queryResultBuffer)];
 		int gpuEnd[arraysize(queryResultBuffer)];
 
+		int gpuBegin_current = -1;
+		int gpuEnd_current = -1;
+
 		bool IsCPURange() const { return !cmd.IsValid(); }
 	};
 	wi::unordered_map<size_t, Range> ranges;
@@ -110,6 +113,7 @@ namespace wi::profiler
 
 		GraphicsDevice* device = wi::graphics::GetDevice();
 		CommandList cmd = device->BeginCommandList();
+		queryheap_idx = device->GetBufferIndex();
 
 		device->QueryReset(
 			&queryHeap,
@@ -130,6 +134,7 @@ namespace wi::profiler
 
 		// note: read the GPU Frame end range manually because it will be on a separate command list than start point:
 		auto& gpu_range = ranges[gpu_frame];
+		gpu_range.gpuEnd_current = gpu_range.gpuEnd[queryheap_idx];
 		gpu_range.gpuEnd[queryheap_idx] = nextQuery.fetch_add(1);
 		device->QueryEnd(&queryHeap, gpu_range.gpuEnd[queryheap_idx], cmd);
 
@@ -147,25 +152,22 @@ namespace wi::profiler
 		);
 
 		nextQuery.store(0);
-		queryheap_idx = (queryheap_idx + 1) % arraysize(queryResultBuffer);
-		uint64_t* queryResults = (uint64_t*)queryResultBuffer[queryheap_idx].mapped_data;
+		const uint64_t* queryResults = (const uint64_t*)queryResultBuffer[queryheap_idx].mapped_data;
 
 		for (auto& x : ranges)
 		{
 			auto& range = x.second;
+			if (!range.in_use)
+				continue;
 
 			if (!range.IsCPURange())
 			{
-				int begin_query = range.gpuBegin[queryheap_idx];
-				int end_query = range.gpuEnd[queryheap_idx];
-				if (queryResultBuffer[queryheap_idx].mapped_data != nullptr && begin_query >= 0 && end_query >= 0)
+				if (queryResults != nullptr && range.gpuBegin_current >= 0 && range.gpuEnd_current >= 0)
 				{
-					uint64_t begin_result = queryResults[begin_query];
-					uint64_t end_result = queryResults[end_query];
+					const uint64_t begin_result = queryResults[range.gpuBegin_current];
+					const uint64_t end_result = queryResults[range.gpuEnd_current];
 					range.time = (float)abs((double)(end_result - begin_result) / gpu_frequency);
 				}
-				range.gpuBegin[queryheap_idx] = -1;
-				range.gpuEnd[queryheap_idx] = -1;
 			}
 			range.times[range.avg_counter++ % arraysize(range.times)] = range.time;
 
@@ -232,8 +234,10 @@ namespace wi::profiler
 		ranges[id].name = name;
 		ranges[id].cmd = cmd;
 
+		GraphicsDevice* device = wi::graphics::GetDevice();
+		ranges[id].gpuBegin_current = ranges[id].gpuBegin[queryheap_idx];
 		ranges[id].gpuBegin[queryheap_idx] = nextQuery.fetch_add(1);
-		wi::graphics::GetDevice()->QueryEnd(&queryHeap, ranges[id].gpuBegin[queryheap_idx], cmd);
+		device->QueryEnd(&queryHeap, ranges[id].gpuBegin[queryheap_idx], cmd);
 
 		lock.unlock();
 
@@ -262,8 +266,10 @@ namespace wi::profiler
 			}
 			else
 			{
+				GraphicsDevice* device = wi::graphics::GetDevice();
+				ranges[id].gpuEnd_current = ranges[id].gpuEnd[queryheap_idx];
 				ranges[id].gpuEnd[queryheap_idx] = nextQuery.fetch_add(1);
-				wi::graphics::GetDevice()->QueryEnd(&queryHeap, it->second.gpuEnd[queryheap_idx], it->second.cmd);
+				device->QueryEnd(&queryHeap, it->second.gpuEnd[queryheap_idx], it->second.cmd);
 			}
 		}
 		else
@@ -330,6 +336,8 @@ namespace wi::profiler
 
 		for (auto& x : ranges)
 		{
+			if (!x.second.in_use)
+				continue;
 			if (x.second.IsCPURange())
 			{
 				if (x.first == cpu_frame)
@@ -354,7 +362,7 @@ namespace wi::profiler
 			{
 				ss << "\t" << x.first << " (" << x.second.num_hits << "x)" << ": " << std::fixed << x.second.total_time << " ms" << std::endl;
 			}
-			else
+			else if(x.second.num_hits == 1)
 			{
 				ss << "\t" << x.first << ": " << std::fixed << x.second.total_time << " ms" << std::endl;
 			}
@@ -371,7 +379,7 @@ namespace wi::profiler
 			{
 				ss << "\t" << x.first << " (" << x.second.num_hits << "x)" << ": " << std::fixed << x.second.total_time << " ms" << std::endl;
 			}
-			else
+			else if (x.second.num_hits == 1)
 			{
 				ss << "\t" << x.first << ": " << std::fixed << x.second.total_time << " ms" << std::endl;
 			}
diff --git a/WickedEngine/wiScene.cpp b/WickedEngine/wiScene.cpp
index ecd2b3078..1621c1eaf 100644
--- a/WickedEngine/wiScene.cpp
+++ b/WickedEngine/wiScene.cpp
@@ -151,7 +151,7 @@ namespace wi::scene
 			}
 
 			// Advance to next query result buffer to use (this will be the oldest one that was written)
-			queryheap_idx = (queryheap_idx + 1) % arraysize(queryResultBuffer);
+			queryheap_idx = device->GetBufferIndex();
 
 			// Clear query allocation state:
 			queryAllocator.store(0);
diff --git a/WickedEngine/wiScene.h b/WickedEngine/wiScene.h
index ee7451e66..3590faf0d 100644
--- a/WickedEngine/wiScene.h
+++ b/WickedEngine/wiScene.h
@@ -125,7 +125,7 @@ namespace wi::scene
 		// Occlusion query state:
 		struct OcclusionResult
 		{
-			int occlusionQueries[wi::graphics::GraphicsDevice::GetBufferCount() + 1];
+			int occlusionQueries[wi::graphics::GraphicsDevice::GetBufferCount()];
 			// occlusion result history bitfield (32 bit->32 frame history)
 			uint32_t occlusionHistory = ~0u;
 
@@ -142,7 +142,7 @@ namespace wi::scene
 		wi::graphics::GPUQueryHeap queryHeap;
 		wi::graphics::GPUBuffer queryResultBuffer[arraysize(OcclusionResult::occlusionQueries)];
 		wi::graphics::GPUBuffer queryPredicationBuffer;
-		int queryheap_idx = 0;
+		uint32_t queryheap_idx = 0;
 		mutable std::atomic<uint32_t> queryAllocator{ 0 };
 
 		// Surfel GI resources:
diff --git a/WickedEngine/wiVersion.cpp b/WickedEngine/wiVersion.cpp
index ebcf623e5..f1ad04253 100644
--- a/WickedEngine/wiVersion.cpp
+++ b/WickedEngine/wiVersion.cpp
@@ -9,7 +9,7 @@ namespace wi::version
 	// minor features, major updates, breaking compatibility changes
 	const int minor = 71;
 	// minor bug fixes, alterations, refactors, updates
-	const int revision = 176;
+	const int revision = 177;
 
 	const std::string version_string = std::to_string(major) + "." + std::to_string(minor) + "." + std::to_string(revision);