Date: Tue, 20 May 2008 18:25:19 -0400
Reply-To: Phil Rack <PhilRack@MINEQUEST.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Phil Rack <PhilRack@MINEQUEST.COM>
Subject: Re: Quad Processors and SAS
In-Reply-To: <200805202220.m4KLE0CO014874@malibu.cc.uga.edu>
Content-Type: text/plain; charset="us-ascii"
Are you running Vista in a Virtual Machine perhaps? Do you have any other
occurrences of SAS running or spawning new jobs? Finally, does this always
happen in the same place in running the program?
Phil
Philip Rack
MineQuest, LLC
SAS & WPS Consulting and Software Development
Tel: (614) 457-3714
-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of P.
Cristian Gugiu
Sent: 05/20/2008 6:20 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: Quad Processors and SAS
I've attempted to exclude every sas affiliated extension. I am still
getting an error message. Here is one of the messages.
Append=91
ERROR: A lock is not available for WORK.VARS.DATA, lock held by another
process.
NOTE: Enter RUN; to continue or QUIT; to end the procedure.
NOTE: Statements not processed because of errors noted above.
NOTE: The SAS System stopped processing this step because of errors.
NOTE: PROCEDURE DATASETS used (Total process time):
real time 0.00 seconds
cpu time 0.00 seconds
Now, there is no reason why the Vars dataset should be locked. That's why
I'm speculating that perhaps SAS or Vista is not coordinating the
processor sharing correctly. I can and hope to be wrong about this. The
offending code appears to be the following.
ods exclude Directory Members ;
proc datasets;
append base=Vars data=V;
run; quit;
I'm creating a temporate dataset V and appending it to my master dataset
Vars. Before this is the IML code that I use to create V. I run this
process iteratively several hundred times. When I get out of testing mode,
I'll probably run the simulation 2000 times. Now, I've thought creating
the entire dataset in one step. My laptop was not able to handle this
load. Some of the datasets have more than 4 million records. Perhaps, my
workstation will be able to handle it. However, I'm not sure if IML can.
My laptop also gave my an error when my matrix had over 650,000 records.
Therefore, I'm not sure if IML can handle matrix algebra when there are
millions of records.
Now, Phil gave me a way to turn off the some of my processros. I turned
off 7 of them and let the program run on just one processor. The program
crashed. I'm starting to wonder if the problem is not Vista. How can the
same program work on an old XP machine but not work on a new Vista
Business machine. Has anyone run into similar issues.
Cristian
