How to use RHI Transition to guarantee the GPU computation is finished?

Development Rendering
,
1

Hello. I am implementing a compute shader. I need to guarantee the computation in GPU has finished before I copy data back to CPU.

Intuitively, I would like to use “fence”, but I found that in UE 5.2 “fence” has been depricated and should be replaced with “transition”. But I do not know how should I use “transition” in UE.

For example, I have a variable and its buffer in my class like this.

TArray<float> Data_Float;
FBufferRHIRef Data_CsBuffer;
FUnorderedAccessViewRHIRef Data_CsBufferUAV;

And I initialize them like this.

		ENQUEUE_RENDER_COMMAND(FComputeShaderRunner)(
			[this](FRHICommandListImmediate& RHICommands)
			{
			TResourceArray<float> TmpResourceArray;
			TmpResourceArray.Init(0, Data_Float.Num());
			FRHIResourceCreateInfo CreateInfo(TEXT("Data_Float"));
			CreateInfo.ResourceArray = &TmpResourceArray;
			constexpr EBufferUsageFlags Usage = BUF_UnorderedAccess | BUF_ShaderResource | BUF_StructuredBuffer;

			Data_CsBuffer = RHICommands.CreateBuffer(
				sizeof(float) * Data_Float.Num(),
				Usage,
				sizeof(float),
				RHIGetDefaultResourceState(Usage, false),
				CreateInfo);
			
			});

After I have triggered some GPU computation, I think I should use a “transition” before I copy back data from Data_CsBufferUAV to Data_Float. I guess it should be something like this.

	RHICommands.Transition(FRHITransitionInfo(Data_CsBufferUAV, ERHIAccess::Unknown, ERHIAccess::UAVMask));

But is this correct? What is the meaning of each ERHIAccess? How and by whom the Data_CsBufferUAV’s state is (shoud should be) changed? Are there any tutorial or documents about these?

Thanks in advance.