Search Results for author:
Found 1 papers, 0 papers with code
TENSILE: A Tensor granularity dynamic GPU memory scheduling method toward multiple dynamic workloads system
In this paper, we demonstrated TENSILE, a method of managing GPU memory in tensor granularity to reduce the GPU memory peak, considering the multiple dynamic workloads.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
