Time sharing massively parallel machines. Draft Page: 3 of 22
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
Time Sharing Massively Parallel Machines
Brent C. Gorda - firstname.lastname@example.org
National Energy Research Supercomputer Center
Lawrence Livermore National Laboratory
Rich Wolski - email@example.com
UCSD Computer Science and Engineering Department
Program development on today's parallel machi a tinues to be difficult due
to lack of good scheduling methodologies. a i development and
production computing are often bes ser n scheduling policies.
The most widely-used solution is to ace ha t e a hi , giving each
user exclusive access to a subset a ava a es rces Unfortunately
since resources are id much o m rin evelof t, space-sharing
can cause the overall mac n be n e u zed. Alternatively,
production use a machin i sp cng environment tends to monopolize
system resou - long pe o f im. The result is poor response time
for interact s
As part of he ye Parallel Computing Initiative (MPCI) at the
Law e\ iv e at'ion Laboratory, we have developed a simple,
fev\a dp time sharing mechanism which addresses these issues
by he n o ocesses on tightly coupled parallel machines. By
time-s irut e rces, our system interleaves production and
interact j s. mmediate priority is given to interactive use,
maintaining response time. Production jobs are scheduled during idle
periods, making use of the otherwise unused resources. In this paper we
discuss our experience with gang scheduling over the 3 year life-time of the
In the next section we motivate the project and discuss some of its details.
Section 3.0 describes the general scheduling problem and how gang scheduling
addresses it. In section 4.0, we describe our implementation. Section 8.0
* Work performed under the auspices of the U. S. Department of Energy by the Lawrence Livermore
National Laboratory under contract No. W-7405-ENG-48.
M A STER TrYUTIOM OF E DOCUMENT IS UNLIMITE Page
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Gorda, B. & Wolski, R. Time sharing massively parallel machines. Draft, article, March 1, 1995; California. (https://digital.library.unt.edu/ark:/67531/metadc624668/m1/3/: accessed March 21, 2019), University of North Texas Libraries, Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.