… In this work, we propose DLPlib, a low-level library for deep learning processors to bridge
the gap between the hardware and the programmers. In DLPLib, all data are represented as …

… Lan [49] and others developed DLPlib as a hardware-oriented library and designed … deep
learning processors for nearly two years, from KNL (Knignts Landing) to Nervana processors […

… DLPlib, a low-level library, provides a series of efficient and versatile programming interfaces
for accelerating various deep learning … for deep learning networks on CPU. Vasilache et al.[…

… , deep learning algorithms are computing intensive and memory intensive. The traditional
way of relying on CPU … -layer neural networks, the reason ZhuQue is faster than DLPlib is that …

… memory usage of neural networks between GPU and CPU memories. The … We use the
upgraded version of DLPlib[32… DLPlib: A library for deep learning processor. Journal of Computer …

… tools, specialized algorithm libraries and visual presentation … The specialized algorithm
libraries make unified invocation … performance computing, deep learning and other customized …

… Martin ( 430 ) Temporally Consistent Depth Map Prediction Using Deep Convolutional Neural
Network and Spatial-Temporal Conditional Random Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . …

… application,the traditional CPU processor is not suitable for … DLPlib 和QingLong 的三 个
版本. 四个对象的区别如表2所示. … is the programming library of the neural network accelerator. …

… 云中的DLPlib 自动完成从Caffe 或TensorFlow 应用到手工程序的变换; 在TensorFlow 对CPU
通用… DLPlib: a library for deep learning processor. J Comput Sci Technol, 2017, 32: 286–296 …

… locality; therefore the deep neural network library has a convention on … this paper, we focus
on deep learning inference. To … For the CPU and GPU platforms we evaluate, we follow the …