@article{tensileatensorgranularitydynamicgpu, title = {TENSILE: A Tensor granularity dynamic GPU memory scheduling method toward multiple dynamic workloads system}, author = {Kaixin Zhang and Hongzhi Wang and Han Hu and Songling Zou and Jiye Qiu and Tongxin Li and Zhishun Wang}, year = {2021}, eprint = {2105.13336}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2105.13336v5}, }