Any particular function to initialize GPU other than the first cudaMalloc call?

by

A call to

cudaFree(0);

is the canonical way to force lazy context establishment in the CUDA runtime. You can’t reduce the overhead, that is a function of driver, runtime and operating system latencies. But the call above will let you control how/when those overheads occur during program execution.

EDIT in 2015 to add that the heuristics of context initialisation in the runtime API have subtly changed over time so that cudaSetDevice now establishes a context, so the cudaFree() call isn’t explicitly required to intialise a context, you can use cudaSetDeviceinstead. Also note that some set-up time will still be incurred at the first kernel launch, whereas before this wasn’t the case. For for kernel timing, it is best to include a warm-up call first before launching the kernel you will time to remove this set-up latency. It appears that the various profiling tools have enough granularity built in to avoid this without any extra API calls or kernel calls.

More Related Contents:

  1. GPU Emulator for CUDA programming without the hardware [closed]
  2. How is CUDA memory managed?
  3. Modifying registry to increase GPU timeout, windows 7
  4. How to measure the inner kernel time in NVIDIA CUDA?
  5. Passing Host Function as a function pointer in __global__ OR __device__ function in CUDA
  6. nvidia-smi Volatile GPU-Utilization explanation?
  7. How do I use Nvidia Multi-process Service (MPS) to run multiple non-MPI CUDA applications?
  8. CUDA apps time out & fail after several seconds – how to work around this?
  9. multi-GPU basic usage
  10. How does CUDA assign device IDs to GPUs?
  11. How do I choose grid and block dimensions for CUDA kernels?
  12. Using Java with Nvidia GPUs (CUDA)
  13. Thrust inside user written kernels
  14. What is the purpose of using multiple “arch” flags in Nvidia’s NVCC compiler?
  15. How are 2D / 3D CUDA blocks divided into warps?
  16. What can I do against ‘CUDA driver version is insufficient for CUDA runtime version’?
  17. What is a bank conflict? (Doing Cuda/OpenCL programming)
  18. CUDA how to get grid, block, thread size and parallalize non square matrix calculation
  19. Can anyone provide sample code demonstrating the use of 16 bit floating point in cuda?
  20. CUDA and nvcc: using the preprocessor to choose between float or double
  21. Polymorphism and derived classes in CUDA / CUDA Thrust
  22. CUDA: How many concurrent threads in total?
  23. How do I select which GPU to run a job on?
  24. How can I make tensorflow run on a GPU with capability 2.x?
  25. How to implement device side CUDA virtual functions?
  26. How to create a CUDA context?
  27. Why does my CUDA kernel crash (unspecified launch failure) with a different dataset size?
  28. CUDA determining threads per block, blocks per grid
  29. How to disable a specific nvcc compiler warnings
  30. “invalid configuration argument ” error for the call of CUDA kernel?

Leave a Comment