LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (April 2011, week 2)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:   Mon, 11 Apr 2011 19:27:14 -0700
Reply-To:   Jack Hamilton <jfh@STANFORDALUMNI.ORG>
Sender:   "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:   Jack Hamilton <jfh@STANFORDALUMNI.ORG>
Subject:   Re: 10 High performance PROC of new SAS
Comments:   To: oloolo <dynamicpanel@YAHOO.COM>
In-Reply-To:   <201104111957.p3BICLEG000486@waikiki.cc.uga.edu>
Content-Type:   text/plain; charset=windows-1252

It's still around, and it's my understanding that it's still under active development.

There are two flavors SPDE server, which runs on its own server as a database, and the SPDE engine, which does not provide all of the benefits of the SPDE server but is part of base.

On Apr 11, 2011, at 12:57 PM, oloolo wrote:

> SAS used to have a technology called SPD, but I am curious how many sites > licensed this engine. > > On Mon, 11 Apr 2011 11:04:08 -0500, Charlie Huang > <charlie.chao.huang@GMAIL.COM> wrote: > >> Oop.. My mistake -- Thank oloolo to correct it. >> >> Parallel processing needs of data partition. The current SAS’s data >> structure, .sas7bdat, does not support it. SAS has to rely on data >> partition from other database system, such as Teradata, to do it. >> >> Those procedures look like well fit SAS’s in-Database technology. >> Revolution R and SAS are on the same road. >> >> Anyway, still not our business, although a HPlogistic procedure looks >> attractive. ….. >> >> Best, >> Charlie >> >> On Mon, Apr 11, 2011 at 10:35 AM, oloolo <dynamicpanel@yahoo.com> wrote: >>> HPMIXED =/= HPLMIXED >>> HPMIXED used sparse matrix techniques to accelerate computing but no > multi- >>> threading/multi-core technology are utilized, and it only deals with a > very >>> restrictied class of linear mixed models, mostly a block-diagonal cov- >>> matrix which is able to use sparse matrix computing to the max power. >>> >>> While I haven't seen any details of HPLMIXED, I believe it is able to >>> handle linear mixed models with richer type of cov-matrix and has mutli- >>> threading/distributed-computing technology embedded. >>> >>> On Mon, 11 Apr 2011 22:26:49 +0800, Murphy Choy <goladin@GMAIL.COM> > wrote: >>> >>>> Woah! Serious? >>>> >>>> Is HPLogistic in SAS/Stat as well? >>>> >>>> It would be great if it is. >>>> >>>> On Mon, Apr 11, 2011 at 10:05 PM, Charlie Huang < >>>> charlie.chao.huang@gmail.com> wrote: >>>> >>>>> HPMIXED is available in SAS/STAT. I guess that others will be licensed >>>>> with SAS/EM, which is irrelevant to most of us. >>>>> >>>>> On Sun, Apr 10, 2011 at 7:47 PM, Murphy Choy <goladin@gmail.com> wrote: >>>>>> HI, >>>>>> >>>>>> From my understanding, many of these procs are in enterprise miner. >>>>>> >>>>>> Not too sure how it will work with the others. >>>>>> >>>>>> On Mon, Apr 11, 2011 at 7:45 AM, oloolo <dynamicpanel@yahoo.com> > wrote: >>>>>> >>>>>>> I read this from SAS blogs, don't know the details, but I am > extremely >>>>>>> interested in all information. So far from the name, I am thinking >>> HPDS2 >>>>>>> may >>>>>>> be a PROC HPDS2, haha >>>>>>> >>>>>>> But since many existing PROC's support programming statements, I > won't >>>>> be >>>>>>> surprised >>>>>>> >>>>>>> On Sun, 10 Apr 2011 18:36:03 -0500, Joe Matise <snoopy369@GMAIL.COM> >>>>>>> wrote: >>>>>>> >>>>>>>> I wonder, if this is going to be a new product (ie, more cost) or >>> part >>>>> of >>>>>>>> 9.3 base (and/or the products these PROCs normally belong to)? >>>>>>>> >>>>>>>> HPDS2 - a proc or a replacement for 'data' (or an addon, like data >>>>>>>> test/hpds2=test; or somesuch)? >>>>>>>> >>>>>>>> -Joe >>>>>>>> >>>>>>>> On Sun, Apr 10, 2011 at 6:25 PM, oloolo <dynamicpanel@yahoo.com> >>>>> wrote: >>>>>>>> >>>>>>>>> Looks very promising, possibly distributed computing mode, >>>>>>>>> >>>>>>>>> The benefits and the PROCS >>>>>>>>> The database appliance provides replication and failover, so that >>> if >>>>> a >>>>>>> node >>>>>>>>> goes down, the database knows where to find it. SAS High- >>> Performance >>>>>>>>> Analytics offers high-end performance analytics with a user >>> interface >>>>>>>>> familiar to SAS users. All the new procedures will work on the >>>>> desktop >>>>>>> as >>>>>>>>> well. >>>>>>>>> >>>>>>>>> Ten new SAS procedures will be offered as part of the release > that >>>>> move >>>>>>>>> existing SAS procedures from single-threaded activities to >>>>>>> multithreaded: >>>>>>>>> >>>>>>>>>   * HPREG: linear regression and variable selection. >>>>>>>>>   * HPLOGISTIC: logistic regression and variable selection. >>>>>>>>>   * HPLMIXED: linear mixed models. >>>>>>>>>   * HPNEURAL: neural nets. >>>>>>>>>   * HPNLIN: nonlinear regression and maximum likelihood. >>>>>>>>>   * HPREDUCE: covariance/correlation analysis, variable > reduction. >>>>>>>>>   * HPDMDB: summarization. >>>>>>>>>   * HPSUMMARY: descriptive statistics. >>>>>>>>>   * HPFOREST: predictive modeling based on decision trees. >>>>>>>>>   * HPDS2: next-generation DATA step. >>>>>>>>> >>>>>>>>> �The new PROCS will be aware of the distributed computer model � >>> You >>>>>>> won�t >>>>>>>>> even notice that it happens,� said Oliver. >>>>>>>>> >>>>>>>>> >>>>>>>>> details see >>>>>>>>> >>>>>>>>> >>>>>>> >>>>>>> >>>>> http://blogs.sas.com/sgf/index.php?/archives/207-The-low-down-on-high- >>> performance-computing-and-10-new-PROCS.html >>>>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Regards, >>>>>> Murphy Choy >>>>>> >>>>>> Certified Advanced Programmer for SAS V9 >>>>>> Certified Basic Programmer for SAS V9 >>>>>> DataShaping Certified SAS Professional >>>>>> >>>>> >>>> >>>> >>>> >>>> -- >>>> Regards, >>>> Murphy Choy >>>> >>>> Certified Advanced Programmer for SAS V9 >>>> Certified Basic Programmer for SAS V9 >>>> DataShaping Certified SAS Professional >>>


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