And regarding point 1 below, I'll raise Andy's $0.02 by a penny or so :)
While servers with large amounts of RAM are now becoming available, the
pipeline architecture and big data library of S-PLUS 7 Enterprise are already
helping many people analyse multi-GB data sets that would swamp the servers
Andy describes. (I'll point you to an example at the end of this message.)
Remember: in versions before S-PLUS 7 Enterprise, you'd often need multiple
copies of a data set in RAM to actually pass the data through an analysis
routine. Even if your data was smaller than the amount of RAM available, it
could still cause the machine to start swapping during the analysis. With the
big data library, the data is not brought into RAM all at once, so this is no
longer an issue. If you're getting out of memory errors with S-PLUS 6 today
(or it's running slowly because the machine is using a lot of virtual memory
and page-swapping frequently), S-PLUS 7 Enterprise will help.
In addition, one of the goals of S-PLUS 7 was to deliver the benefits of
statistical analysis across organizations, especially to those that are not
statisticians themselves. (For more information about what I mean by this,
see: http://www.insightful.com/insightful_doclib/document.asp?id=168.) This
includes statistical analysis of large data sets. And while a single
multi-GB server might be affordable, it's unlikely to be affordable to
purchase one for each of dozens or hundreds of users. One of the great
benefits of S-PLUS 7 Enterprise is that it allows the analysis of large data
sets with hardware already in place, even on modest desktop systems that do
not have large amounts of RAM.
And finally, while the big data library of S-PLUS 7 enables analysis of large
data sets, it also adds new capabilities to the S language that apply to both
small and large data sets, like data preparation tools, SQL data processing,
and new graphical techniques like Trellised hexbin plots.
If you'd like to see an example of the value these capabilities of S-PLUS 7
Enterprise provide, Greg Gupton of Moody's will be giving a webinar on April
2, showing how he uses the big data library to analyze large data sets and
create some incredible graphics of the results using the new techniques I
mentioned. Details are available at:
http://www.insightful.com/news_events/webcasts/2005/08gupton/default.asp
# David Smith
-----Original Message-----
From: Liaw, Andy [mailto:andy_liaw@merck.com]
Sent: Tue 7/12/2005 9:08 PM
To: 'Thomas Jagger'; s-news@lists.biostat.wustl.edu
Cc:
Subject: Re: [S] Concerns about SPLUS version 7
I'll chip in my $0.02:
1. I find Insightful's marketing blurb for the big data library
in the Enterprise Developer version a bit puzzling: I just
check the price of a dual Opteron 275 server with 8GB of RAM
at one vendor, and it is under $6500 (with 16GB that jumps
to over $12000). Anyone care to disclose the price tag of
the enterprise developer edition?
|