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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The all-to-all collective communications primitive is widely used in machinelearning (ML) and high performance computing (HPC) workloads, and optimizingits performance is of interest to both ML and HPC communities. All-to-all is aparticularly challenging workload that can severely strain the underlyinginterconnect bandwidth at scale. This is mainly because of the quadraticscaling in the number of messages that must be simultaneously serviced combinedwith large message sizes. This paper takes a holistic approach to optimize theperformance of all-to-all collective communications on supercomputer-scaledirect-connect interconnects. We address several algorithmic and practicalchallenges in developing efficient and bandwidth-optimal all-to-all schedulesfor any topology, lowering the schedules to various backends and fabrics thatmay or may not expose additional forwarding bandwidth, establishing an upperbound on all-to-all throughput, and exploring novel topologies that delivernear-optimal all-to-all performance.

Efficient All-to-All Collective Communication Schedules for Direct-Connect Topologies