'Use all 24 Gb for one application on the K80 GPU

The Tesla K80 has 24 Gb but, as far as I understand, it is shared between two GK210 GPUs on the same card. So actually it's a card with two 12 Gb GPUs. But is it still possible to use all 24 Gb for one application fx training a large model in PyTorch or Keras?

keraspytorch


Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source

Related Questions

Validation loss and loss stuck at 0

Create custom convolution layer and compare two keras layers

Is it possible to continue training from a specific epoch?

Error: ValueError: The last dimension of the inputs to `Dense` should be defined. Found `None`

Sparse Categorical CrossEntropy shape problem with Keras

input_shape of Conv1D layer Keras

How to load data from a downloaded tar.gz file in tensorflow/keras?

Specifying the batch size when subclassing keras.model

Tensorflow - ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type float)

Whenever i try to print classification report, in accuracy column it prints some other value 0.50 but my accuracy is 0.96

Keras verbose training progress bar writing a new line on each batch issue

Keras verbose training progress bar writing a new line on each batch issue

Could not load dynamic library 'cudart64_101.dll' on tensorflow CPU-only installation

Keras - Plot training, validation and test set accuracy

Keras: network doesn't train with fit_generator()