Date: Thu, 30 Apr 2009 08:53:34 -0400
Reply-To: Michael Raithel <michaelraithel@WESTAT.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Michael Raithel <michaelraithel@WESTAT.COM>
Subject: Re: Optimal Hardware Configuration -- white paper?
Content-Type: text/plain; charset="us-ascii"
Kristin Graves posted the following:
> Hello --
> My company is running some vendor-supplied SAS code. We've
> done test runs on our PC/network versus the vendor's. The
> program runs in about four hours on their site but takes
> several days on our site.
> The vendor mentioned that there is a SAS white paper that
> discusses how your hardware can be optimized to speed up SAS
> runs. Would anyone happen to have a link to that white paper?
> We are running SAS 9.1.3 on a Windows XP Pro operating system.
Kristin, I see from your follow-up that you received some good input both from oloolo on the 'L and from some private sources. So, you are probably along your way to crafting an approach to your investigation. I wanted to add a few words of my own.
First, I definitely recommend both of Tony Brown's papers; even thought the second one is more UNIX than Windows oriented. He is fast becoming one of my favorite SAS Institute authors.
Secondly, I would question the results that your vendor is claiming to get when running the program. Are they _REALLY_ running the program in four (4) hours? How do you know that is true? Do you have a SAS log, or are they just telling you that it took four hours to run? Establishing a clear baseline run time is vital to your own efforts to match or beat the vendor's run time. So, I would double check the assertion that it ran in four hours.
Thirdly, I would ask the vendor about their computer configuration. So many things are important:
1. The make, model, speed, and memory of the computer the SAS program was run on
2. The RAID configuration of the disk drives the data and SAS work area were on
3. Whether the data was local to the computer running the SAS program or on a network drive
4. Whether the SAS work area was local to the computer running the SAS program or on a network drive
5. If data or the SAS work area was on a network drive, what was the network traffic like on the day it was run
6. If the program was run on a shared server (e.g. Windows 2003), how many other tasks were running simultaneously
7. What version of SAS was being run
8. Other hardware/software issues that I just can't think of right here; right now, but that some SAS-L smarty will chime in with
Fourthly, I would ask my own staff... questions 1 - 8, above, and compare the results to what the vendor stated.
Unfortunately, I think that this is far from a simple issue and will take a bit of work to get sorted out.
Best of luck in all of your SAS endeavors!
I hope that this suggestion proves helpful now, and in the future!
Of course, all of these opinions and insights are my own, and do not reflect those of my organization or my associates. All SAS code and/or methodologies specified in this posting are for illustrative purposes only and no warranty is stated or implied as to their accuracy or applicability. People deciding to use information in this posting do so at their own risk.
Michael A. Raithel
"The man who wrote the book on performance"
Author: Tuning SAS Applications in the MVS Environment
Author: Tuning SAS Applications in the OS/390 and z/OS Environments, Second Edition
Author: The Complete Guide to SAS Indexes
A painting in a museum hears more ridiculous opinions than
anything else in the world. - Edmond de Goncourt