How to decide ThreadPerBlock in customised cuda kernel?

Hi,
I’m using AGX Xavier, and NX Xavier with customized cuda kernel.
I’m trying to understand how to decide <<<block, thread >>> value?

for example, I’m using 3840 x 2160 image and AGX Xavier (https://developer.ridgerun.com/wiki/index.php?title=Xavier/JetPack_4.1/Components/Cuda#:~:text=Max%20dimension%20size,1024%2C%201024%2C%2064))

in this case, how do you decide the value?

thank you.

Hi @jahwan.oh, the maximum number of threads per block is one of the GPU specifications in CUDA, you can print them with deviceQuery utility:

Device 0: "Xavier"
  CUDA Driver Version / Runtime Version          10.2 / 10.2
  CUDA Capability Major/Minor version number:    7.2
  Total amount of global memory:                 31929 MBytes (33479647232 bytes)
  ( 8) Multiprocessors, ( 64) CUDA Cores/MP:     512 CUDA Cores
  GPU Max Clock rate:                            1377 MHz (1.38 GHz)
  Memory Clock rate:                             1377 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 524288 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            Yes
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 0 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 10.2, CUDA Runtime Version = 10.2, NumDevs = 1
Result = PASS

So for AGX Xavier, it’s Maximum number of threads per block: 1024 (or 32x32)

The reason that I use a smaller value of 8x8 is this code supports older Tegra devices going back to TX1.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.