About: Scalable Hierarchical Aggregation and Reduction Protocol (SHARP)(TM) Streaming-Aggregation Hardware Design and Evaluation

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Scalable Hierarchical Aggregation and Reduction Protocol (SHARP)(TM) Streaming-Aggregation Hardware Design and Evaluation Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	Scalable Hierarchical Aggregation and Reduction Protocol (SHARP)(TM) Streaming-Aggregation Hardware Design and Evaluation
Creator	Cho, David Klein, Daniel Bloch, Gil Burredy, Devendar Elias, George Graham, Richard Ladd, Joshua Levi, Lion Maor, Ophir Marelli, Ami Petrov, Valentin Qin, Yong Romlet, Evyatar Shainer, Gilad Zemah, Ido
Source	PMC
abstract	This paper describes the new hardware-based streaming-aggregation capability added to Mellanox’s Scalable Hierarchical Aggregation and Reduction Protocol in its HDR InfiniBand switches. For large messages, this capability is designed to achieve reduction bandwidths similar to those of point-to-point messages of the same size, and complements the latency-optimized low-latency aggregation reduction capabilities, aimed at small data reductions. MPI_Allreduce() bandwidth measured on an HDR InfiniBand based system achieves about 95% of network bandwidth. For medium and large data reduction this also improves the reduction bandwidth by a factor of 2–5 relative to host-based (e.g., software-based) reduction algorithms. Using this capability also increased DL-Poly and PyTorch application performance by as much as 4% and 18%, respectively. This paper describes SHARP Streaming-Aggregation hardware architecture and a set of synthetic and application benchmarks used to study this new reduction capability, and the range of data sizes for which Streaming-Aggregation performs better than the low-latency aggregation algorithm.
has issue date	2020-05-22(xsd:dateTime)
bibo:doi	10.1007/978-3-030-50743-5_3
has license	cc-by
sha1sum (hex)	3ed10468b420b300a23737e958a8aad5624554d0
schema:url	https://doi.org/10.1007/978-3-030-50743-5_3
resource representing a document's title	Scalable Hierarchical Aggregation and Reduction Protocol (SHARP)(TM) Streaming-Aggregation Hardware Design and Evaluation
has PubMed Central identifier	PMC7295336
schema:publication	High Performance Computing
resource representing a document's body	covid:3ed10468b420b300a23737e958a8aad5624554d0#body_text
is schema:about of	named entity 'PyTorch' named entity 'medium' named entity 'network bandwidth' covid:arg/3ed10468b420b300a23737e958a8aad5624554d0 named entity 'MPI' named entity 'algorithms' named entity 'low-latency' named entity 'InfiniBand' named entity 'This' named entity 'Mellanox' named entity 'HDR' named entity 'bandwidths' named entity 'Reduction' named entity 'low-latency' named entity 'bandwidth' named entity 'bandwidth' named entity 'software-based' named entity 'algorithm' named entity 'InfiniBand' named entity 'MPI' named entity 'Hardware Design' named entity 'PCI bus' named entity 'algorithm' named entity 'ping-pong' named entity 'radix' named entity 'MPI' named entity 'unicast' named entity 'PCIe' named entity 'ping-pong' named entity 'orders of magnitude' named entity 'low-latency' named entity 'High Performance Computing' named entity 'Mellanox' named entity 'vector' named entity 'data throughput' named entity 'vector' named entity 'low-latency' named entity 'multicast' named entity 'data distribution' named entity 'InfiniBand' named entity 'Mellanox' named entity 'vectors' named entity 'Mellanox' named entity 'collective operation' named entity 'short message' named entity 'PCIe' named entity 'point-to-point' named entity 'throughput' named entity 'PCI bus' named entity 'CPU' named entity 'Mellanox' named entity 'InfiniBand' named entity 'message size' named entity 'firmware' named entity 'MTU' named entity 'latency' named entity 'wire speed' named entity 'asynchronous' named entity 'bandwidth' named entity 'low-latency' named entity 'half and half' named entity 'latency' named entity 'sec' named entity 'radix' named entity 'network bandwidth' named entity 'electrostatics' named entity 'network bandwidth' named entity 'vectors' named entity 'MByte'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software