Paper accepted at SC2024
Congratulations to all co-authors (Munkyu Lee, Sihoon Seong, Minki Kang, Jihyuk Lee, Gap-Joo Na, In-Geol Chun, Cheol-Ho Hong and Dimitrios S. Nikolopoulos), for a paper on ParvaGPU: Efficient Spatial GPU Sharing for Large-Scale DNN Inference in Cloud Environments, accepted at SC2024. We present rigorous algorithms and a system for reducing the GPU footprint of large-scale inference worklaods.