# ATC 2022 ## Meta Info 2022 USENIX Annual Technical Conference Homepage: Paper list: ## Papers ### DNN Inference * [DVABatch: Diversity-aware Multi-Entry Multi-Exit Batching for Efficient Processing of DNN Services on GPUs](https://www.usenix.org/conference/atc22/presentation/cui) \[[Personal Note](/reading-notes/conference/atc-2022/dvabatch.md), [Code](https://github.com/sjtu-epcc/DVABatch)] * [Serving Heterogeneous Machine Learning Models on Multi-GPU Servers with Spatio-Temporal Sharing](https://www.usenix.org/conference/atc22/presentation/choi-seungbeom) \[[Personal Note](/reading-notes/conference/atc-2022/gpulet.md), [Code](https://github.com/casys-kaist/glet)] * [SOTER: Guarding Black-box Inference for General Neural Networks at the Edge](https://www.usenix.org/conference/atc22/presentation/shen) \[[Personal Note](/reading-notes/conference/atc-2022/soter.md), [Code](https://github.com/hku-systems/SOTER)] ### DNN Training * [Whale: Efficient Giant Model Training over Heterogeneous GPUs](https://www.usenix.org/conference/atc22/presentation/jia-xianyan) \[[Personal Note](/reading-notes/conference/atc-2022/whale.md), [Code](https://github.com/alibaba/EasyParallelLibrary)] ### Resource Manager * [PilotFish: Harvesting Free Cycles of Cloud Gaming with Deep Learning Training](https://www.usenix.org/conference/atc22/presentation/zhang-wei) \[[Personal Note](/reading-notes/conference/atc-2022/pilotfish.md)] * [Memory Harvesting in Multi-GPU Systems with Hierarchical Unified Virtual Memory](https://www.usenix.org/conference/atc22/presentation/choi-sangjin) \[[Personal Note](/reading-notes/conference/atc-2022/memharvester.md)] --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://paper.lingyunyang.com/reading-notes/conference/atc-2022.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.