In this letter, we present deep partitioned training to accelerate computations involved in training DNN models. This is the first work that partitions a DNN model across storage devices, an NPU and a host CPU forming a unified compute node for training workloads. To validate the benefit of using the proposed system during DNN training, a trace-based simulator or an FPGA prototype is used to estimate the overall performance and obtain the layer index to be partitioned that provides the minimum latency. As a case study, we select two benchmarks, i.e., vision-related tasks and a recommendation system. As a result, the training time reduces by 12.2 ~ 31.0 percent with four near-storage computing devices in vision-related tasks with a mini-batch size of 512 and 40.6 ~ 44.7 percent with one near-storage computing device in the selected recommendation system with a mini-batch size of 64.

中文翻译：

---

从近存储计算到 DNN 加速器的深度分区训练


在这封信中，我们提出了深度分区训练，以加速训练 DNN 模型所涉及的计算。这是第一个将 DNN 模型跨存储设备、NPU 和主机 CPU 分区的工作，形成用于训练工作负载的统一计算节点。为了验证在 DNN 训练期间使用所提出的系统的好处，使用基于轨迹的模拟器或 FPGA 原型来估计整体性能并获得提供最小延迟的要分区的层索引。作为案例研究，我们选择两个基准，即视觉相关任务和推荐系统。结果，在小批量大小为512的视觉相关任务中，使用四台近存储计算设备时，训练时间减少了12.2%~31.0%；在所选推荐系统中，使用一台近存储计算设备时，训练时间减少了40.6%~44.7%。小批量大小为 64。