How to allocate GPU memory in DRIVE AGX PEGASUS only with iGPU

I’m evaluating DRIVE AGX PEGASUS (only with iGPU) in the viewpoint of performance and determinism.

Would you introduce which document explains the following items?

  1. Do Tegra CPU and Tegra iGPU share the same system LPDDR4?

  2. If yes for No. 1, when does the GPU driver allocate/deallocate the memory?
    E.g., allocate at system startup and never deallocate until system shutdown.

Dear kenji.funaoka,
Both CPU and iGPU share same physical memory. When cudaMalloc() gets called it allocates memory on GPU. When an application allocates memory on GPU, the de allocation happens as soon as application terminates. Please check https://docs.nvidia.com/cuda/cuda-for-tegra-appnote/index.html#memory-management for more details on memory architecture.

Thank you SivaRamaKrishna! My question was solved.