|
The Palo Alto,
California-based company said the prototype contains 160
terabytes of memory, capable of managing the information from
every book in the U.S. Library of Congress five times over.
It is the latest prototype from "The Machine" research project
by HPE, which aims to create super-fast computers by designing
them around memory. Traditionally, the way processors, storage
and memory interact can bog down computers.
The prototype underscores HPE's ambition to lead computer
technology as huge datasets place new strains on devices.
"We need a computer built for the Big Data era," HPE's Chief
Executive Meg Whitman said in a news release.
While large data centers that piece together many computers may
have enough calculating power, they cannot transfer data
efficiently, said Kirk Bresniker, Hewlett Packard Labs Chief
Architect, in an interview. That means HPE's single-system model
may one day compete with the infrastructure spearheaded by
cloud-computing companies like Amazon.com Inc.
HPE expects its model will over time contain more and more
memory. While the prototype remains years away from being
commercially available, HPE is already bringing some of the tech
from its research program to market.
Still, companies and the scientific community have yet to concur
on what technology will best serve users.
"You need computing that scales up with the size of the
dataset," said Kathy Yelick, a professor of electrical
engineering and computer sciences at the University of
California at Berkeley.
There's still discussion "about what the right answer is."
(Reporting by Jeffrey Dastin in San Francisco; Editing by
Stephen Coates)
[© 2017 Thomson Reuters. All rights
reserved.] Copyright 2017 Reuters. All rights reserved. This material may not be published,
broadcast, rewritten or redistributed.
 |
|
