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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

While the jury is still out whether Moore&#x0027;s law is dead or not, it is undeniable that the performance improvement of general purpose processors has slowed down in the last decade. This has caused the computing industry to revisit the customization and specialization fronts, and advance approaches based on heterogeneous accelerators. The emergence of deep learning as a important workload has further accelerated the rapid appearance of new computing &#x201C;species&#x201D;. This is similar to what happened during the Cambrian explosion 500 million years ago, a sudden diversification of life forms which changed the biosphere forever. However, for an ecological structure to become sustainable, the rest of the environment also has to change. This talks argues that today&#x0027;s computing environment is not yet ready to fully embrace accelerators and heterogeneity. New interconnects, communication protocols, and mechanisms to share data need to be redesigned from the ground up for the new breed of accelerators to thrive. Industry-wide efforts promoting new interconnect standards could provide the necessary support, are promising, but have not yet matured. As an example, the talk concludes with an overview of how HPE is using one of these protocols, Gen-Z, to develop a high-end memory-driven computing approach that can lead to efficient, balanced systems at exascale and beyond.

Computing in the Cambrian Era Plenary Talk