|Date: ||Thu, 14 Sep 2000 14:31:26 -0400|
|Reply-To: ||"Muhlbaier, Lawrence H." <lawrence.muhlbaier@DUKE.EDU>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||"Muhlbaier, Lawrence H." <lawrence.muhlbaier@DUKE.EDU>|
|Organization: ||Duke Clinical Research Institute|
|Subject: ||Re: SAS - Unix vs NT in batch|
|Content-Type: ||text/plain; charset=us-ascii|
SAS in batch works fine on NT. I actually run the same programs both places
without changing any code save my libname/filename type statements. Some
caveats for large datasets:
--NT is likely slower than Unix here. No formal tests, just a feel.
--If you do not store the data local, you will be using a lot of time just
moving data across the network. Consider the NT server version of SAS.
--The 'at' command in NT is a bit different. In particular, it requires
administrator level rights to use, so our network admins won't allow us to use
it. There are third party software products that provide the unix-like 'at' job
scheduling features (though I haven't tried any).
--File compression is available at the operating system level on NT (no more
gzip'ing), but I have had some problems with > 2GB files.
"Fry, Steve" wrote:
> My agency presently runs SAS V8 on Solaris -Unix OS in a client-server
> architecture. We process very large datasets (i.e. multigig, millions of
> obs) always in batch mode. There is some discussion here about migrating to
> NT. I have zero experience with running SAS on NT. Question: Is running SAS
> in similar fashion even feasible on an NT platform?
Lawrence H. ('Doc') Muhlbaier email@example.com
Assistant Research Professor
Duke University Medical Center 919-668-8774 (office)
DUMC 3865 919-383-0595 (home)
Durham, NC 27710-7510 919-668-7057 (FAX)