English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

At the heart of deep learning training and inferencing are computationallyintensive primitives such as convolutions which form the building blocks ofdeep neural networks. Researchers have taken two distinct approaches tocreating high performance implementations of deep learning kernels, namely, 1)library development exemplified by Intel MKL-DNN for CPUs, 2) automaticcompilation represented by the TensorFlow XLA compiler. The two approaches havetheir drawbacks: even though a custom built library can deliver very goodperformance, the cost and time of development of the library can be high.Automatic compilation of kernels is attractive but in practice, till date,automatically generated implementations lag expert coded kernels in performanceby orders of magnitude.  In this paper, we develop a hybrid solution to the development of deeplearning kernels that achieves the best of both worlds: the expert codedmicrokernels are utilized for the innermost loops of kernels and we use theadvanced polyhedral technology to automatically tune the outer loops forperformance. We design a novel polyhedral model based data reuse algorithm tooptimize the outer loops of the kernel. Through experimental evaluation on animportant class of deep learning primitives namely convolutions, we demonstratethat the approach we develop attains the same levels of performance as IntelMKL-DNN, a hand coded deep learning library.

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