Tango: A Deep Neural Network Benchmark Suite for Various Accelerators
Tango: A Deep Neural Network Benchmark Suite for Various Accelerators
复制标题
Tango：适用于各种加速器的深度神经网络基准套件
DOI：
10.1109/ispass.2019.00021
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Hyeran Jeon
中科院分区：
文献类型：
--
作者：
A. Karki;Chethan Palangotu Keshava;Spoorthi Mysore Shivakumar;Joshua Skow;Goutam Madhukeshwar Hegde;Hyeran Jeon
深度神经网络（DNN）已在多个计算领域证明了其有效性。为给DNN应用提供更高效的计算平台，拥有包含各种基准工作负载的评估环境至关重要。尽管最近已经发布了一些DNN基准测试套件，但其中大多数都需要安装专有DNN库或资源密集型DNN框架，这些在资源有限的移动平台或架构模拟器上难以运行。为提供一个更具扩展性的评估环境，我们提出了一个新的DNN基准测试套件，它可以在任何支持CUDA和OpenCL的平台上运行。所提出的基准测试套件包括最广泛使用的五种卷积神经网络和两种循环神经网络。我们提供了这些网络在架构模拟器、服务器GPU和移动GPU以及移动FPGA上运行时的架构统计信息。
Deep neural networks (DNNs) have been proving the effectiveness in various computing fields. To provide more efficient computing platforms for DNN applications, it is essential to have evaluation environments that include assorted benchmark workloads. Though a few DNN benchmark suites have been recently released, most of them require to install proprietary DNN libraries or resource-intensive DNN frameworks, which are hard to run on resource-limited mobile platforms or architecture simulators. To provide a more scalable evaluation environment, we propose a new DNN benchmark suite that can run on any platform that supports CUDA and OpenCL. The proposed benchmark suite includes the most widely used five convolution neural networks and two recurrent neural networks. We provide architectural statistics of these networks while running them on an architecture simulator, a server- and a mobile-GPU, and a mobile FPGA.