Home / ubuntu / Questions / 1142735
In Process
Kong
Asked: 2019-05-13 16:15:02 +0800 CST

How to understand CUDA out of memory message?

  • 772

Below is the output of nvidia-smi.

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.104      Driver Version: 410.104      CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 108...  On   | 00000000:01:00.0 Off |                  N/A |
| 57%   68C    P8    19W / 280W |   7278MiB / 11176MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX 108...  On   | 00000000:02:00.0 Off |                  N/A |
| 67%   73C    P2   259W / 280W |  10873MiB / 11178MiB |     96%      Default |
+-------------------------------+----------------------+----------------------+

Now I try to run a code and it gives me this message.

RuntimeError: CUDA out of memory. Tried to allocate 8.62 MiB (GPU 0; 10.91 GiB total capacity; 2.80 GiB already allocated; 16.88 MiB free; 0 bytes cached)

I understand that I do not have enough memory but where do I see how much memory is required by my code?

I try to run another code that requires x10000 more memory and it gives me this error

RuntimeError: CUDA out of memory. Tried to allocate 858.38 MiB (GPU 0; 10.91 GiB total capacity; 2.51 GiB already allocated; 756.88 MiB free; 0 bytes cached)

How can I make it tell me something like CUDA out of memory. Need 12GB Memory total.

cuda

0 Answers