LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous (more recent) messageNext (less recent) messagePrevious (more recent) in topicNext (less recent) in topicPrevious (more recent) by same authorNext (less recent) by same authorPrevious page (August 1998, week 2)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Wed, 12 Aug 1998 15:16:59 -0700
Reply-To:     "Self, Karsten" <kself@VISA.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Self, Karsten" <kself@VISA.COM>
Subject:      Re: SAS and RAM disks
Comments: To: "Bassett Consulting Services, Inc."
          <BASSETT.CONSULTING/0002395748@mcimail.com>
Content-Type: text/plain; charset="iso-8859-1"

I don't think this is quite correct.

While it's true that caching loads hard drive contents to RAM, and that creating a dedicated RAMdrive makes less memory available to other operations, it might still make sense to do this if you are on a multi-user or multi-tasking system with enough I/O activity to effectively ensure that the information you'd like to have cached from disk has been flushed from the cache by the time you access it again. While you could attempt to solve the problem by throwing RAM at your box until caching did produce the desired level of performance, you could get more predictable results by dedicating a portion of memory to the same purpose.

I'm flying a bit by the seat of my pants on this. I'm not sure whether/when a RAMdrive might find itself swapped out to VM, or what affect different choices of OS would have on performance.

My choice: I probably wouldn't set up a RAMdrive for small files accessed frequently which would probably find themselves in cache anyway. I probably would set up a RAMdrive for middlin' to large files, accessed less frequently, but for which high performance is important. More memory on the box would allow greater discretion in allocating resources for a dedicated RAMdisk.

Moral: try it and see what your results are.

-- Karsten M. Self (kself@visa.com) Trilogy Consulting

What part of "Gestalt" don't you understand?

> ---------- > From: Bassett Consulting Services, > Inc.[SMTP:BASSETT.CONSULTING/0002395748@mcimail.com] > Sent: Tuesday, August 11, 1998 8:38 PM > To: SAS-L@UGA.CC.UGA.EDU > Subject: Re: SAS and RAM disks > > ---------------------------------------------------------------------- > - > CONTENT: re: SAS and RAM disks > NAME: Michael L. Davis > INTERNET: Bassett.Consulting@worldnet.att.net > AFFILIATION: Bassett Consulting Services, Inc. > P-ADDR: 10 Pleasant Drive, North Haven, CT 06473 > PHONE: (203) 562-0640 > FAX: (203) 498-1414 > ---------------------------------------------------------------------- > - > > A couple of messages were posted on SAS-L earlier today, asking about > the wisdom of setting up a RAM disk under Windows 95 to speed up SAS. > Great thinking but probably unnecessary. My understanding is that > Windows 95, NT and other recent releases of computer operating cache > the most recent pages of disk files accessed to RAM. > > So for practical purposes, if you have enough RAM so that significant > portions of your disk files can be cached, you are automatically are > getting the speed advantages that previously required a RAM disk to > implement. I also understand that some high performance disk drives > may have their own RAM for caching that is independent of the RAM > memory elsewhere in the computer. > > By contrast, if you resort to a RAM disk and something bad happens, > you run a high risk of losing the data processed with no hope of > recovering it. So please make sure that if you use a RAM disk that > you faithfully save the contents to a less volatile form of storage! > > As a shameless plug, consultants really interested in tuning their > PC systems and those of their clients to better run SAS should > consider > attending the CONSUG meeting scheduled for Sunday morning of the > NESUG '98 conference. One of the featured speakers, Peter Ruzsa is > to address this subject. For more information, visit the NESUG 98 > web site at http://www.pace.edu/nesug/98/ > > -------------------------- original posting ------------------------ > > Date: Tue Aug 11, 1998 2:49 pm EST > From: "Raymond V. Liedka" > EMS: INTERNET / MCI ID: 376-5414 > MBX: liedka@UNM.EDU > > TO: SAS-L > EMS: INTERNET / MCI ID: 376-5414 > MBX: SAS-L@UGA.CC.UGA.EDU > Subject: Re: SAS and RAMdisk > > On Tue, 11 Aug 1998, SPRUCK, Dirk CEN/US wrote: > > > I am currently working with SAS 6.12 under Windows 95. Since my PC > is > > equipped with a lot of RAM I am thinking about installing a RAMdisk > and > > direct the SAS working directory to that RAMdisk. Does anyone have > > experience with that? Will it make SAS performing faster? Is is > worth > > the effort? I know that RAMdisks are considered "old technology" > (which > > I don't understand), but why not using it if it speeds up SAS? > > Dirk, > > I asked the same question to our department computer nerd geek techie, > and > he kind of turned up his nose at the idea. His argument is that the > newest hard disk drives are really quite fast and that the gain in > speed > by using a RAMdisk probably wouldn't amount to anything. > > However, I am now going to be getting a new computer with 192 MB of > RAM, > and it sure seems to me that setting up (only for SAS work) a RAMdisk > of > 100 MB or so would greatly speed up some work I am planning this fall > involving simulations of data sets with several million cases and > repeated > sorts and random draws of those data sets. > > Anyway...I don't have the answer, but if you recieve a private > response, I > would greatly appreciate you passing any info on. > > Thanks, > > ray > > Raymond V. Liedka > Assistant Professor > Department of Sociology > University of New Mexico >


Back to: Top of message | Previous page | Main SAS-L page