Deep learning (DL) frameworks take advantage of GPUs to improve the speed of DL inference and training. Ideally, DL frameworks should be able to fully utilize the computation power of GPUs such that the running time depends on the amount of computation assigned to GPUs. Yet, we observe that in scheduling GPU tasks, existing DL frameworks suffer from inefficiencies such as large scheduling overhead and unnecessary serial execution. To this end, we propose Nimble, a DL execution engine that runs GPU tasks in parallel with minimal scheduling overhead. Nimble introduces a novel technique called ahead-of-time (AoT) scheduling.

Nimble introduces a novel technique called ahead-of-time (AoT) scheduling.

Nimble introduces a novel technique called ahead-of-time (AoT) scheduling.

Modern deep neural network (DNN) training jobs use complex and heterogeneous software/hardware stacks. The efficacy of software-level optimizations can vary significantly when used in different deployment configurations. It is onerous and error-prone for ML practitioners and system developers to implement each optimization separately, and determine which ones will improve performance in their own configurations. Unfortunately, existing profiling tools do not aim to answer predictive questions such as "How will optimization X affect the performance of my model?". We address this critical limitation, and proposes a new profiling tool, Daydream, to help programmers efficiently explore the efficacy of DNN optimizations. Daydream models DNN execution with a fine-grained dependency graph based on low-level traces collected by CUPTI, and predicts runtime by simulating execution based on the dependency graph. Daydream maps the low-level traces using DNN domain-specific knowledge, and introduces a set of graph-transformation primitives that can easily model a wide variety of optimizations. We show that Daydream is able to model most mainstream DNN optimization techniques, and accurately predict the efficacy of optimizations that will result in significant performance improvements.

In Personalization at Amazon, we use neural networks to generate personalized product recommendations for our customers. Amazon's product catalog is huge compared to the number of products that a customer has purchased, making our datasets extremely sparse. And with hundreds of millions of customers and products, our neural network models often have to be distributed across multiple GPUs to meet space and time constraints. For this reason, we have created and open-sourced DSSTNE, the Deep Scalable Sparse Tensor Neural Engine, which runs entirely on the GPU. We use DSSTNE to train neural networks and generate recommendations that power various personalized experiences on the retail website and Amazon devices.