krj@cs.toronto.edu (Ken Jackson) (02/01/90)
========================================================================
() _ _ _ __ _
/\ ' ) / ' ) ) // _/_ _/_
/ ) _. o ( / o / / _ , , , _ // _ / / _ __
/__/__(__<_ \/ <_ / (_</_(_(_/_/_)_</_</_<___<__</_/ (_
========================================================================
Postings can be submitted by sending mail to sci-vi@acm.princeton.edu.
Additions and deletions should be sent to sci-vi-request@acm.princeton.edu.
Issues of this digest is available by anonymous ftp from acm.princeton.edu.
Wed Jan 3 15:21:17 1990
Volume 01: Issue 01
Today's Topics:
Administrivia
another animation system
cache memory traces
Wanted: more info on apE 2.0
your request for vis software info (comp.graphics)
----------------------------------------------------------------------
Date: Wed, 3 Jan 90 12:09:09 -0500
From: markv (Mark VandeWettering)
Subject: Administrivia
To: sci-vi
Welcome to the first real edition of the Sci Vi News. I have been playing
around with some mail digesting software in an effort to take the drudgery
out of dealing with mail, and am now moderately pleased with the results.
The limiting speed of the time between receiving submissions and forwarding
them should not be the speed of your moderator, but rather the speed at
which submissions are generated.
All submissions to the group should go to sci-vi@acm.princeton.edu. Requests
for additions and deletions can go to sci-vi-request@acm.princeton.edu.
DO NOT send mail to me directly for info or requests about the Sci Vi
News, I have a separate mailbox for these messages. (I am not going to
even mention my address, in an effort to keep people who might
otherwise not know it from knowing it).
There are now 191 people on this list (some duplicates I think, cure that
by sending me a message at sci-vi-request), with vastly varying interests
and experience. My original intention was to publish all the bibliographies
that were more than a single line, but that would make a truly unwieldy
issue. If their is sufficient interest, I will generate such a "who's who"
issue at a later date.
------------------------------
Date: Mon, 4 Dec 89 22:13:52 EST
From: mccalpin@masig3.ocean.fsu.edu (John D. McCalpin)
Subject: another animation system
To: markv@acm
At Florida State, we use an animation package developed locally.
It is called `lookat', and is (currently) a three-pass system.
(1) Run the graphics program on the mainframe (or whatever).
This program calls routines very similar to the NCAR graphics
library, but with some additions for animation.
The output is a device independent ASCII metacode.
(2) Move the metacode to the workstation (usually an SGI IRIS)
and translate (using `mtran') into a binary, random-access file.
(3) Run `lookat' to control the display of the animated movie.
`lookat' controls which frames to show and in what order, what
overlays to show on each frame, which color table to use, and
what location to look at the image from.
We are considering combining passes (1) and (2), since we now have
the capability of making binary files of the desired structure on
all of our machines.
For more info on the distribution, contact Al Davis, comp. software spec.,
davis@masig1.ocean.fsu.edu
------------------------------
Date: Fri, 29 Dec 89 13:41:37 PST
From: mdeale@cosmos.acs.calpoly.edu (Myron Deale)
Subject: cache memory traces
To: sci-vi@acm
Hello gang,
to start with, I'll say that my problem is not well defined.
I'm thinking of trying to visualize a typical microprocessor
instruction (and possibly data) address trace. Consider a 32-bit
machine with its 4 Billion word address space. And then I want
to try and see patterns in the way the micro accesses memory
-- over time.
I want to view the data on a Sun 3/60 color monitor,
1152 x 900 x 256. Yes, I've read the paper by what's-his-name
(Voldheim ?) at IBM [one of the co-authors was Mandelbrot]. I
suppose I could use their run-length encoding and extract fractal
dimensions, but I don't want Yet Another Senior Project. I'd like
to start with some reasonably simple data reduction tricks ....
The computational load could be distributed among our network
of Sun's, since the address trace file has already been generated.
The trace file contains about 2 million addresses, resides on the
server (naturally) and could easily be divied up and sent to various
clients. I think.
Note that I'm not particularly interested in computing the hit
rate of a cache memory. I'm interested in viewing addressing patterns
and sequences. I'll intuit from there.
-Myron
// mdeale@cosmos.acs.calpoly.edu
[ Moderator -- If you are just interested in the qualitative behavior of
the instruction traces, then try some relatively simple things. You have
something on the order of 1Kx1K display area. You need to display the
memory behavior of a process, which probably doesn't exceed 1 megabyte of
code space, it almost certainly doesn't exceed 16 megabytes.
As an instruction is executed at a particular address, you might wish
to "flash" it, and then let the colors decay at some rate. Color mapping
hardware could be useful in this. Inner loops would then occur as brighter
regions on the screen if displayed fast enough.
To compress the data, a variant of run-length encoding would probably
be good for instruction streams, where sequential accesses are run-length
encoded together.
This is an interesting project, but I am tempted to say something like
"Unless you know the question, what can you learn from the answer?"
The more specific the question that you wish to address, the better
your attempts at finding the answers to visualization will be.
Mark VandeWettering ]
------------------------------
Date: Wed, 3 Jan 90 15:14:34 EST
From: michelle@rhett.osgp.osc.edu (Michelle Messenger)
Subject: Wanted: more info on apE 2.0
To: markv@acm
Hi Mark,
The following information is what we normally send out to people who have
questions about the apE. If you would like additional information for your
newsletter, please let me know.
Thanks for your interest in the apE!
--Michelle Messenger
Project Coordinator
Ohio Supercomputer Graphics Project
1224 Kinnear Road
Columbus, Ohio 43212
614-292-3274
michelle@rhett.osgp.osc.edu
********************************************************************************
Facts About the apE Release:
Here are some details about the availability of the apE software:
What is It?
The apE is a software toolkit for visualizing scientific data. It is designed
to provide the flexibility to examine data in several ways. Release 1.1 of
the apE is designed primarily for visualizing data in two graphical
dimensions. (this does not mean just for visualizing two-dimensional data).
The apE system places an emphasis on animation and the integration of the
visualization process into the scientific process.
The apE is extendable, and will grow with the addition of more tools and
filters. The apE is a software product resulting from leading research in
computer graphics and scientific visualization conducted by The Ohio
Supercomputer Graphics Project and the Advanced Computing Center for the
Arts and Design. The apE 2.0, available in winter 1990, will address
general visualization problems in both two and three dimensions.
What Computers is it For?
The software package is designed so that there are two types of programs:
interactive tools, and computational filters. The tools allow for interactive
manipulation and examination of data. The filters perform all the computations.
The filters are usable without the tools, although the user may need to
provide some programming support to create data flows and view rendered images.
The complete apE toolkit, including the tools (running under SunView),
filters, and documentation, is available for the following workstations:
- Sun series 4 workstations.
- Sun series 3 workstations, running SunOS version 4.0 or greater.
- Sun series 386i workstations.
How Do I Get it?
The apE is available via anonymous ftp from oscsuna.osc.edu (128.146.1.4) in
the directory pub/apE. Send electronic mail to michelle@rhett.osgp.osc.edu,
for more information.
What Does the Release Include?
- Executable versions of all the programs
(tools and filters) for Sun-4, Sun-3, and Sun 386i.
- Complete user documentation for each program.
- Examples and tutorial.
Source code for the apE version 1.1 is not available; however, future
releases will include source code at a nominal fee (<$100).
apE Version 2.0
The apE 2.0 represents a significant release of graphics software and tools
to the research community. Built upon the experience of staff members from
the Ohio Supercomputer Graphics Projects and the Advanced Computing Center
for the Arts and Design, the apE 2.0 includes software that addresses the
full range of applications, from science to art.
The apE 2.0 is projected for release in January, 1990, and will be available
in binary form via anonymous ftp much like the apE 1.1. However, full
source code and printed documentation (both user and programmer) will be
available as well for a nominal fee. The apE 2.0 will include libraries,
filters, tools, and applications. In addition, a much larger range of
equipment, including HP, DEC, SGI, and others, will be supported
under the X Window system.
The apE 2.0 will include basis libraries, libraries for manipulating data flow
structures, and higher level libraries for window system objects (browsers,
etc.). In addition to the two dimensional tools for scientific data included
in the 1.1 release, the apE 2.0 will also include a complete polygonal system,
with data generation, animation, and rendering; several mapping systems from
scientific data to surface data; a volumetric rendering system; full image
manipulation and post processing software; a primitive based rendering system
(sticks, ball, vectors, etc.), ray tracing software, and a full toolset for
scientific and artistic image creation and manipulation.
------------------------------
Date: Mon, 4 Dec 89 21:13:34 MST
From: rasure@borris.unm.edu (John Rasure)
Subject: your request for vis software info (comp.graphics)
To: sci-vi@acm
Here at the University of New Mexico, we developed a system called
XVision. It is primarily for image processing. It has been licensed
by Paragon Imaging and here is a blurb.
John Rasure
-----------------------------
The Visualization Workbench (TM)
University pricing has been set as:
Image Processing Library and
Visual Programming System $500.00
Image Processing Library Source $500.00
Plotting Package $500.00
The above prices area available for single user copies to
Universities, for more pricing information and details
please contact:
Paragon Imaging
171 Lincoln Street
Lowell, MA 01852
508-441-2112 (phone)
508-459-9719 (FAX)
paragon@bullwinkle.unm.edu
(please send all e-mail to this address)
The Visualization Workbench (TM)
--- ------------- ---------
The Visualization Workbench (TM) is an application oriented,
X Windows based (MIT), electronic imaging software system.
It integrates image processing with graphics, visual pro-
gramming and automatic generation of graphical user inter-
face code. It enables application developers and research-
ers to quickly analyze imaging problems, prototype solutions
and develop application programs. For OEM product
designers, the Visualization Workbench is a system on which
to build more competitive products and reduce product
development time.
The Visualization Workbench passes other systems in
five principal areas:
o completeness
o user acceptance
o portability and networking
o open architecture
o application-specific adaptation without programming
Completeness
------------
The Visualization Workbench is an integrated general
purpose electronic imaging system designed for user exten-
sion and application customization. Features include:
o three user interfaces
o automatic generation of X Windows user interfaces
o instruction and learning aides
o an extensive image algorithm library
o automatic generation of program skeletons for new algo-
rithms
o interactive 2D and 3D image plotting and display
including the visualization of user defined mathemati-
cal expressions
o interactive image display and manipulation
o hardcopy output
o open image file structure
User Acceptance
---- ----------
The Visualization Workbench includes three user inter-
faces to meet user and application preferences:
o a visual programming system
o menu/form interface
o command line user interface
In addition, a journal/playback system can be used with the
user interfaces for demonstration and instruction.
Portability and Networking
----------- --- ----------
The Visualization Workbench is designed to operate in a
heterogeneous networked environment. It is based on
standards - UNIX, X Windows and Postscript.
Open Architecture
---- ------------
The user interface tools and image file structure allow
seamless interfaces to other applications.
Application Specific Adaptation Without Programming
----------- -------- ---------- ------- -----------
Through a powerful set of tools the Visualization Work-
bench can be expanded and customized to application specific
requirements without programming.
The Visual Programming System - Paragon Pipes
The innovative visual programming system allows users
to program a complete image processing application through
object manipulation. To develop a program the user interac-
tively selects, via mouse, the imaging algorithms which con-
stitute the processing task. Each function is displayed as
a "glyph", a symbolic representation of the function.
The user creates processing pipes by visually connect-
ing the glyphs in the desired order. Within a pipe, the
user can interactively view data and manipulate the program
parameters, and efficiently ask what-if questions.
Application Specific User Interfaces
The Visualization Workbench allows users to create new
user interfaces for any imaging application. By editing a
user interface specification via the graphical interactive
editor or any text editor, the user can modify an existing
user interface or generate a new one. This include the
visual programming system as well as the menu/form user
interface.
Interactive Plotting
Users can interactively generate and display 2D and 3D
plots of images, mathematical functions and data points.
One or more plots can be viewed interactively from any per-
spective. 2D plot types supported are: line, discrete,
histogram and scatter; 3D plot types supported are: line,
mesh, scatter, and contour. The plots can be freely anno-
tated.
Interactive Image Display
The user can interactively display and process images
of any size. Interactive functions include zoom and pan,
overlays, region of interest processing (any algorithm),
pseudo coloring and pixel display.
Image Algorithm Library
The 120 algorithms are categorized as follows:
o color imaging and transformations
o arithmetic and logic functions
o geometric operations
o classification of multiband images
o FFT and correlation
o frequency domain filtering
o spectral display
o image enhancement
o image creation
o spatial domain processing
o statistical processing
o segmentation
o feature extraction
o image dithering
The Visualization Workbench is currently supported on
DEC Stations, Vaxen, Sun 3, SUN 4, and Sparc. Sony News, Concurrent,
Data General and Ardent Titan are also supported.
------------------------------
End of Sci-Vi Digest
******************************
Reposted by
Prof. Kenneth R. Jackson, krj@na.toronto.edu (on Internet, CSNet,
Computer Science Dept., ARPAnet, BITNET)
University of Toronto, krj@na.utoronto.ca (on CDNnet and other
Toronto, Ontario, X.400 nets (Europe))
Canada M5S 1A4 ...!{uunet,pyramid,watmath,ubc-cs}!utai!krj
(Phone: 416-978-7075) (on UUCP)
(FAX: 416-978-4765)