TESLA PCIe V100 32G + Z10PE-D16 WS Unboxing-Part 3(GPU/CPU simple test)

Following the parts of hardware and operating system installation in the previous article, we will unpack the GPU test process today.

1. Install nvidia-cuda-toolkit before testing (2,007 MB of disk space):

sudo apt install nvidia-cuda-toolkit

2. check nvcc version:

3. If you cannot see the installed version of nvcc here, first, check whether there is nvcc in the bin directory of cuda:

4. If yes, add the path to the system path, edit bashrc:

nano ~/.bashrc

5. Add the following text and save it:

export PATH=/usr/local/cuda-10.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH

6. Note that if you have different CUDA directories. Be sure to specify the directory that holds the nvcc file:

7. The above steps may fail. If it cannot be solved, please install the driver manually:

8. After that, find the corresponding CUDA version of the driver and download it.

Start installing:

9. After completing the installation, check the card status:

10. If the installed version is not correct, it may cause problems with the operation of your system. (or even not working.)

11. The above steps complete the driver and CUDA TOOLS, then install CuDNN.

Download URL:

https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/

installation:

sudo dpkg -i libcudnn7_7.6.5.32–1+cuda10.2_amd64.deb

Note that it must match your CUDA version

12. Install TensorRT

sudo apt-key add /var/nvinfer-runtime-trt-repo-5.0.2-ga-cuda10.0/7fa2af80.pub

sudo dpkg -i nvinfer-runtime-trt-repo-ubuntu1804–5.0.2-ga-cuda10.0_1–1_amd64.deb

13. Install anaconda

sudo apt install libgl1-mesa-glx libegl1-mesa libxrandr2 libxrandr2 libxss1 libxcursor1 libxcomposite1 libasound2 libxi6 libxtst6

wget -P /tmp https://repo.anaconda.com/archive/Anaconda3-2020.02-Linux-x86_64.sh

sha256sum /tmp/Anaconda3–2020.02-Linux-x86_64.sh

bash /tmp/Anaconda3–2020.02-Linux-x86_64.sh

source ~/.bashrc

14. Create a virtual environment

We are determined to build a python version 3.6 to correspond to the old version of tensorflow.

conda create -n gputest36 python=3.6

15. Start the virtual environment

conda activate gputest36

16. Install tensorflow and other related packages for testing

pip install tensorflow-gpu==1.15
pip install keras
pip install matplotlib
pip install pandas
pip install jupyter
python -m ipykernel install — user — name gputest36

17. Start jupyter notebook

18. To confirm whether Tensorflow works with GPU

19. Simply test the CPU/GPU performance

/gpu:0 size= 0 Time: 0.000536727905273
/cpu:0 size= 0 Time: 0.06526756286621094
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 1000 Time: 0.07515811920166016
/cpu:0 size= 1000 Time: 0.14328646659851074
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 2000 Time: 0.017438650131225586
/cpu:0 size= 2000 Time: 0.0789041519165039
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 3000 Time: 0.017871618270874023
/cpu:0 size= 3000 Time: 0.21936392784118652
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 4000 Time: 0.040592193603515625
/cpu:0 size= 4000 Time: 0.45297980308532715
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 5000 Time: 0.03450465202331543
/cpu:0 size= 5000 Time: 0.6990058422088623
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 6000 Time: 0.04923081398010254
/cpu:0 size= 6000 Time: 1.1312360763549805
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 7000 Time: 0.07115960121154785
/cpu:0 size= 7000 Time: 1.858928918838501
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 8000 Time: 0.0996253490447998
/cpu:0 size= 8000 Time: 2.8014204502105713
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 9000 Time: 0.13735628128051758
/cpu:0 size= 9000 Time: 3.7534072399139404
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 10000 Time: 0.17871403694152832
/cpu:0 size= 10000 Time: 4.951266288757324
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 11000 Time: 0.22140908241271973
/cpu:0 size= 11000 Time: 6.701673984527588
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 12000 Time: 0.28014397621154785
/cpu:0 size= 12000 Time: 8.391860961914062
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 13000 Time: 0.361297607421875
/cpu:0 size= 13000 Time: 10.389846086502075
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 14000 Time: 0.44246649742126465
/cpu:0 size= 14000 Time: 13.142382621765137
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 15000 Time: 0.5497150421142578
/cpu:0 size= 15000 Time: 15.8779938220977787
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 16000 Time: 0.6581192016601562
/cpu:0 size= 16000 Time: 23.798622846603394
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 17000 Time: 0.7908139228820801
/cpu:0 size= 17000 Time: 28.801256895065308
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 18000 Time: 0.9442639350891113
/cpu:0 size= 18000 Time: 33.42823314666748
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 19000 Time: 1.1140697002410889
/cpu:0 size= 19000 Time: 39.92689752578735
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 20000 Time: 1.2889599800109863
/cpu:0 size= 20000 Time: 47.23131489753723
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 21000 Time: 1.507587194442749
/cpu:0 size= 21000 Time: 53.957911252975464
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 22000 Time: 1.7296056747436523
/cpu:0 size= 22000 Time: 62.29582238197327
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 23000 Time: 1.9872767925262451
/cpu:0 size= 23000 Time: 70.74891448020935
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 24000 Time: 2.540881872177124
/cpu:0 size= 24000 Time: 80.50912141799927
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 25000 Time: 8.453091621398926
/cpu:0 size= 25000 Time: 92.09209561347961
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 26000 Time: 31.823472499847412
/cpu:0 size= 26000 Time: 105.20583033561707
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 27000 Time: 35.68541669845581
/cpu:0 size= 27000 Time: 117.7775731086731
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 28000 Time: 39.85190725326538
/cpu:0 size= 28000 Time: 133.0733766555786
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 29000 Time: 44.32763409614563
/cpu:0 size= 29000 Time: 142.72973489761353
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 30000 Time: 49.07072424888611
/cpu:0 size= 30000 Time: 156.02221703529358
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 31000 Time: 54.23987174034119
/cpu:0 size= 31000 Time: 175.54739713668823
— — — — — — — — — — — — — — — — — — —
/gpu:0 size= 32000 Time: 91.86028575897217
/cpu:0 size= 32000 Time: 192.83744859695435
— — — — — — — — — — — — — — — — — — —

20. Draw the result

YOLOV5 to be continued.