Back to Results
First PageMeta Content
Supercomputers / Computer buses / Local area networks / Message Passing Interface / QsNet / Remote direct memory access / InfiniBand / Quadrics / Myrinet / Computing / Parallel computing / Computer networks


Optimizing Bandwidth Limited Problems Using One-Sided Communication and Overlap Christian Bell1,2 , Dan Bonachea1 , Rajesh Nishtala1 , Katherine Yelick1,2 {csbell,bonachea,rajeshn,yelick}@cs.berkeley.edu Computer Science
Add to Reading List

Document Date: 2006-04-20 04:11:42


Open Document

File Size: 288,31 KB

Share Result on Facebook

Company

IBM / Mellanox / Optimizing Bandwidth Limited / supNetwork Hardware / /

/

Event

Natural Disaster / /

Facility

GASNet communication system / University of California / /

IndustryTerm

higher overall efficiency on systems / unordered network hardware / shared memory hardware / scientific applications / cross-processor / interconnect technologies / bandwidth-sensitive applications / eager algorithm / host processor / interconnect solutions / distributed algorithm / real-world scientific applications / communication protocol / terconnect hardware / lower software / parallel systems / production applications / cluster networks / communication startup / target networks / cluster network hardware / given network / target processor / lower software overhead / cluster hardware / processor-based solutions / computing / communication network / remote processors / near-commodity network / rendezvous algorithm / /

MarketIndex

MPI / NAS FT / NAS / /

OperatingSystem

UNIX-like / /

Organization

University of California / Berkeley1 Computational Research Division / UPC’s SPMD / Computer Science Division / /

Person

Lawrence Berkeley / /

Position

translator / repeated forward / representative for one-sided communication / programmer / /

Product

GASNet / One-Sided Communication / /

ProgrammingLanguage

Fortran / /

ProvinceOrState

California / /

RadioStation

Core / /

Technology

Alpha / 3-D / UNIX / API / NIC processor / implementation algorithms / eager algorithm / UDP / RDMA-based interconnect technologies / shared memory / operating systems / increasingly offloading communication protocol / Ethernet / target processor / rendezvous algorithm / host processor / Pencils algorithm / MPI protocol / cache-blocked algorithm / 512 processors / /

SocialTag