You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A major performance bottleneck in the current implementation is the use of a dictionary to store path tensors in the class PathData. This creates a huge performance overhead when using CUDA vs. CPU., as each individual tensor in the dictionary must be copied to the GPU separately.
A better solution could possibly be implemented based on the new nested tensor class available in pyTorch 2.1. This would allow to store all paths with varying lengths in a single nested tensor, i.e. we would only need a single copy operation when moving the data to the GPU.
The text was updated successfully, but these errors were encountered:
For a test implementation in which I replaced the dictionary in PathData by a nested_tensor the performance of the PathData creation on the GPU increases by a factor of five.
The unfold operation, which we need to create edge_indices of k-th order models is not yet implemented for nested_tensors (neither for the CPU nor for GPU backend)
A major performance bottleneck in the current implementation is the use of a dictionary to store path tensors in the class
PathData
. This creates a huge performance overhead when using CUDA vs. CPU., as each individual tensor in the dictionary must be copied to the GPU separately.A better solution could possibly be implemented based on the new nested tensor class available in pyTorch 2.1. This would allow to store all paths with varying lengths in a single nested tensor, i.e. we would only need a single copy operation when moving the data to the GPU.
The text was updated successfully, but these errors were encountered: