Applying the Roofline model for Deep Learning performance optimizations

09/23/2020

∙

In this paper We present a methodology for creating Roofline models automatically for Non-Unified Memory Access (NUMA) using Intel Xeon as an example. Finally, we present an evaluation of highly efficient deep learning primitives as implemented in the Intel oneDNN Library.

READ FULL TEXT

Applying the Roofline model for Deep Learning performance optimizations

Intel nGraph: An Intermediate Representation, Compiler, and Executor for Deep Learning

PolyScientist: Automatic Loop Transformations Combined with Microkernels for Optimization of Deep Learning Primitives

Softmax Optimizations for Intel Xeon Processor-based Platforms

Deep Learning Based Cryptographic Primitive Classification

Evaluation of the Intel Xeon Phi and NVIDIA K80 as accelerators for two-dimensional panel codes

Intel RealSense Stereoscopic Depth Cameras

Studying the Potential of Automatic Optimizations in the Intel FPGA SDK for OpenCL

Applying the Roofline model for Deep Learning performance optimizations

Related Research

Intel nGraph: An Intermediate Representation, Compiler, and Executor for Deep Learning

PolyScientist: Automatic Loop Transformations Combined with Microkernels for Optimization of Deep Learning Primitives

Softmax Optimizations for Intel Xeon Processor-based Platforms

Deep Learning Based Cryptographic Primitive Classification

Evaluation of the Intel Xeon Phi and NVIDIA K80 as accelerators for two-dimensional panel codes

Intel RealSense Stereoscopic Depth Cameras

Studying the Potential of Automatic Optimizations in the Intel FPGA SDK for OpenCL