From davidb@mcnc.org (David Bennett)
Subject: AVS '94 Registration/Hotel Forms
Message-ID: <1994Feb1.154212.23176@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: robin.mcnc.org
Organization: MCNC Center for Microelectronics, RTP, NC
Date: Tue, 1 Feb 1994 15:42:12 GMT

1994 INTERNATIONAL AVS USERS CONFERENCE REGISTRATION FORM 
Please type or print clearly. Photocopy additional forms if 
necessary. 

FULL NAME___________________________________
TITLE_______________________________________
COMPANY____________________________________
ADDRESS_____________________________________
____________________________________________
CITY___________________	STATE______
ZIP/POSTAL CODE_________	COUNTRY_______
TELEPHONE _____________	FAX______________

CONFERENCE FEES:	ADVANCE REGULAR AMOUNT
(Postmarked			Before 		After
				3/18/94)		3/18/94
Commercial Attendees:
AVS User Group Member		$525			$625 __________
Non Member			$575			$675 __________
Student/Government:*
AVS User Group Member		$425			$525 __________
Non Member			$475			$575 __________

* Must include a copy of current identification or documentation of 
your student or government status with this form 

You must be a member of the IAC PRIOR tooApril 22, 1994 to 
receive the $50 discount.  Membership is $36 from the IAC.

Training Sessions (Please specify which session(s) you enroll in)

IMPORTING DATA INTO AVS	May 1 (am)
WRITING AVS MODULES	May 3 (am)		May 4 (am)
AVS SPECIAL TOPICS - TIPS AND TRICKS 
May 1 (pm) 	May 2 (pm)		May 3 (pm)
Training Sessions	#______	@ $75 per session	________

TOTAL CONFERENCE REGISTRATION ENCLOSED			__________

PAYMENT OPTIONS:
Please return this form with full payment or it cannot be processed. 
Mail or fax your registration form with credit card payments. Fees 
are also payable in U.S. Funds by Check or money order. Checks or 
money orders must made payable to Advanced Visual Systems Inc. 
By Mail:

The 1994 AVS User Group Conference 
c/o J.R. Schuman Associates 
800 South Street, Suite 255
Waltham, MA 02154 USA

By Fax: You may fax this form to J.R. Schuman Associates at (617) 
235-5560:

(Please check one):	Master Card	Visa
CARD NUMBER__________________________________________ 
EXPIRATION DATE________________________________________ 
SIGNATURE_____________________________________________ 
PRINT NAME OF CARDHOLDER________________________________ 

Confirmation:
All registration forms received by April 15th will receive a written 
confirmation. After April 15th, you may contact the AVS Conference
Registration line at (617) 891-8406 if you wish to	confirm.

Conference Registration Cancellations/Substitutions 
All cancellations and substitutions must be made in writing and 
mailed or faxed to:

1994 AVS User Group Conference	
c/o J.R. Schuman Associates
800 South Street, Suite 255
Waltham, MA 02154 U.S.A.
FAX: (617)235-5560

Cancellations must be received by April 15th in writing and will 
receive a 100% refund. There will be no refunds after April 15. All 
refunds will be sent after the conclusion of the Conference. 
Substitutions may be made at anytime.



	1994 INTERNATIONAL AVS USERS CONFERENCE 
			HOTEL REGISTRATION FORM

THIS FORM MUST BE SENT DIRECTLY TO THE SHERATON BOSTON HOTEL 
& TOWERS To assure proper room registration for your stay. please 
complete this reservation request and return before APRIL 11, 1994. 
Requests received after this date will be accepted based on room 
and rate availability. All reservation requests must be accompanied 
by a ONE NIGHT ROOM deposit plus 9.7% tax by check, or credit cards 
listed below. Requests received without a one night deposit or 
guarantee will not be honored. Non-guaranteed reservations are 
subject to cancellation if not guaranteed prior to arrival.

NAME: ________________________________________________ 
ADDRESS:______________________________________________ 
CITY:	____________________ 
STATE:_________________________
ZIP/POSTAL CODE:________________ 
PHONE:________________________ 
ARRIVAL DATE:__________________..
DEPARTURE DATE:__________________ 
SHARING WITH:__________________________________________
SHERATON CLUB INTERNATIONAL # (if applicable):_______________

ACCOMMODATIONS:				RATE	TOTALS
Daily Rate for Single Or Double		$ 140	_________
Government Rate for Single or Double 	$ 90*	_________

*Current and valid Identification must be presented to the hotel in 
order to receive this	special Government rate.

Extra Person Charge Per Room		$ 20	__________
SUB TOTAL__________
Sales Tax:	9.7% subject to change		__________
TOTAL ONE NIGHT DEPOSIT 			__________

NOTE:
All hotel accounts are payable at departure, subject to prior credit 
arrangements at time of registration.

Check in time will be after 3 p.m. on date of arrival. Check out time 
is 12:00 p.m.

No charge for children under 18 years of age when sharing room with 
parent in existing bedding.

I would like to receive rental and catering information on reserving 
a hospitality suite.

I desire a wheelchair accessible room.

I prefer a non-smoking room

ONE NIGHT DEPOSIT:
Check for first night deposit and tax enclosed, OR Charge my 
American Express card, Visa, MasterCard, Diners Club or Discover 
Card	for first night's deposit and tax.

I will use this card to settle my account upon departure
YES		NO
CREDIT CARD TYPE: (Please check one)
AMEX	VISA	MC	DINER's	DISCOVER
CREDIT CARD # _______________________ 
EXP. DATE___________ 
CARDHOLDERS AUTHORIZED SIGNATURE:______________________ 

MAIL THIS FORM TO :	
Sheraton Boston Hotel & Towers
39 Dalton St.
Boston, MA 02199
HOTEL REGISTRATION 
BY FAX:		617-236-6095
BY PHONE:	617-236-2000

Please indicate your affiliation with the AVS Conference when 
making your reservation in order to receive the special conference 
hotel rate. 

-- 
David Bennett
Information Technology Division, MCNC
248 1182


From coullet@sucre.unice.fr (Pierre Coullet)
Newsgroups: comp.graphics.avs
Subject: very naive question
Date: 28 Jan 1994 16:54:28 GMT
Organization: University of Nice Sophia-Antipolis
Lines: 2
Message-ID: <2ibg04$gmk@taloa.unice.fr>
NNTP-Posting-Host: sucre.unice.fr

What is the name of the company in fance which delivers AVS
Thanks


From ferguson@craycos.com (Scott Ferguson)
Newsgroups: comp.graphics.avs
Subject: Up Vector for Planet Earth
Date: 28 Jan 1994 15:24:07 -0700
Organization: Cray Computer Corporation
Lines: 22
Distribution: world
Message-ID: <2ic3a7$7qj@conq2.craycos.com>
NNTP-Posting-Host: conq2.craycos.com

I've been playing with the world coordinates (get it? The geometry
of the continents, countries, etc from the ftp site (avs.ncsc.org)), as well
as the field_to_globe module.

When you do this, you come to realize that the north Pole is at Z = 1 for the
polyline data for the countries, etc, while the north Pole is at Y = 1 in the
field_to_globe module. This means you've gotta do a 2-axis rotation to line
them up, rotating by 270 degrees about X, then by 270 degrees around Y (if
you're rotating the polyline data).

Which is the more common usage? I'd like to try to stay in one coordinate
system, since rotating the earth off its axis could have dire consequenses...
also in case I start playing with some other public domain earth data I'd
like to minimize the amount of transforming I have to do.

Thanks,
Scott
-- 
------------------------------------------------------------------------------
Scott Ferguson                                      Cray Computer Corporation
ferguson@craycos.com                                Colorado Springs, CO 80906
------------------------------------------------------------------------------


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - add_2_ints
Message-ID: <1994Feb3.214839.8173@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 3 Feb 1994 21:48:39 GMT

Name        : add_2_ints      Version      : 1.000     Mod Number : 1769
Author      : Steve Thorpe, International AVS Center
Submitted   : 02/03/94        Last Updated : 02/03/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/add_2_ints
Ported to   : Kubota Sun HP
Description : This short AVS module was written to add 2 integers
              together. When creating an animation using
              animate_integer to 1) drive orthogonal slicer's slice
              plane, and 2) same integer to drive an output frame name (as
              created using Terry Myerson's animate_file), I wanted an
              offset to be added to the file name's frame number. This
              offset is provided by a parameter to this module.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - Field_Math
Message-ID: <1994Feb3.214911.8240@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 3 Feb 1994 21:49:11 GMT

Name        : Field_Math      Version      : 1.000     Mod Number : 1770
Author      : Krzysztof (Chris) S. Nowinski, Institute of Applied
              Mathematics and Mechanics, Warsaw University
Submitted   : 02/03/94        Last Updated : 02/03/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/Field_Math
Ported to   : Sun HP
Description : The module does some field mathematics -- I am intended to
              talk about it at the AVS94. The module interprets a short (up
              to 60 lines) program written in AVSFOL (A Very Simple Field
              Operation Language). AVSSPL admits two types of variables
              - simple variables with values being single float numbers;
              field variables with values being float arrays of
              dimensions determined by the output field.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From rjt@ornl.gov (Ross J. Toedte)
Subject: integration with Alias/Wavefront
Message-ID: <rjt-020294080646@rjtmac.ctd.ornl.gov>
Followup-To: comp.graphics.avs
Sender: usenet@ornl.gov (News poster)
Organization: Oak Ridge National Laboratory
Date: Wed, 2 Feb 1994 13:06:46 GMT
Lines: 8

  We have considerable experience using AVS to build 3D visualizations.
Now, we want to be able to use AVS for developing the visual paradigms and
add value with one of the high-end modelling/animation/rendering packages
such as Wavefront Advanced Visualizer or Alias. I'd like to hear how other
people are integrating the functionality of packages such as these with
visualization toolkits such as AVS, Iris Explorer, Data Explorer, et al.

Ross


From mayalama@willis.cis.uab.edu (Mallik Yalamanchili)
Subject: Converting a ps file to a geom format.
Message-ID: <1994Feb2.163532.11770@cis.uab.edu>
Sender: mayalama@cis.uab.edu (Mallik Yalamanchili)
Organization: University of Alabama at Birmingham
Date: Wed, 2 Feb 1994 16:35:32 GMT
Lines: 8

Hi,
I am trying to convert a post-script file to a geometry format so that it can
be read into AVS networks using 'read geom' module. Could somebody suggest me
the procedure to do that? I could not find any 'direct' modules at IAC to do
that. If anyone has a routine/module that does that or has a network of modules
that does that, pls email me. Any pointers and suggestions are appreciated.
Thanks in advance.
-- Mallik.


From rsignell@crusty.er.usgs.gov (Richard P. Signell)
Subject: AVS benchmarking (How does YOUR system rate?)
Message-ID: <CKM4ty.Kwr@netnews.whoi.edu>
Sender: news@netnews.whoi.edu
Nntp-Posting-Host: 128.128.40.19
Organization: U.S. Geological Survey, Woods Hole, MA
Date: Wed, 2 Feb 1994 19:58:45 GMT
Lines: 53

I'm about to buy a new machine, and I'm wondering what
the hottest machines are for running AVS.  I've tested
several different machines on some simple AVS benchmarks, but
I'd love to see what *YOUR* machine does on them.  I'd be very
interested in hearing from people with AVS running on 
Convex, SUN, HP and IBM systems.  I'll be happy to collect e-mail
responses and post a summary to this newsgroup.   

Here's what I did:

Graphics Benchmarks with AVS-

Test 1: Teapot. Invoke avs with "avs -timer".  
Bring up the geometry viewer.
Load the "teapot.geom" object from the /usr/avs/data/geometry directory.  
Start it spinning.  Stop it.
In the window you invoked AVS, you will see the number of
frames per second (FPS) that your machine produced, and gives
you a solid number on 3D graphics performance to compare with 
other machines running AVS.

Test 2: Mandrill on Teapot.  In the geometry viewer, click on 
"edit texture", then select the "mandrill.x" image from the 
/usr/avs/data/image directory.  Click on "set dynamic texture".
(on machines without hardware texture mapping, you will need to 
change the "camera" setting from "hardware renderer" to "software renderer".
Start it spinning. Stop it.  The FPS this time gives you a solid number
of how well your machine does on 3D graphics performance with 
texture mapping.


Here's what I found:

Tested machines to date (ranked in order of 3D performance):

                             3D Performance   Texture-Mapped 3D Performance
                              Teapot (FPS)     Mandrill on Teapot (FPS)  
Kubota 3400-E210                  72                 10.4
SGI Crimson VGXT, R4000           60                 15.0
SGI Indigo2-Extreme,R4000         52                  1.7
SGI Indigo ELAN, R4000            30                  1.9
Dec 5000/200 PXG/TURBO            18                  0.5 
SGI Indigo XS-24, R3000           12                  0.8


Please help me add to this list!

Thanks,
--
Rich Signell               |  rsignell@crusty.er.usgs.gov
U.S. Geological Survey     |  (508) 457-2229  |  FAX (508) 457-2310
Quissett Campus            |  "What you don't know CAN hurt you,
Woods Hole, MA  02543-1598 |    only you won't know it. "


From billc@avs.com (William G. Clements)
Newsgroups: comp.graphics.avs
Subject: X11R5, SunOS4 GS board
Date: 2 Feb 1994 22:23:10 GMT
Organization: Advanced Visual Systems Inc.
Lines: 7
Message-ID: <2ip94e$5bb@nda.nda.com>
NNTP-Posting-Host: mercury.avs.com
X-Newsreader: TIN [version 1.2 PL1]

Anybody running AVS on the above config?  There appears to be no
cg12.c in the MIT X11R5 distribution so where did you get the
GS driver??

Thanks much,

Bill


From lloyd@silvaco.com (Lloyd Evans)
Subject: AVS benchmarking
Message-ID: <CKMKts.M11@silvaco.com>
Sender: usenet@silvaco.com
Nntp-Posting-Host: post
Reply-To: lloyd@silvaco.com
Organization: Silvaco International.
Date: Thu, 3 Feb 1994 01:44:16 GMT
Lines: 48


Rich> Graphics Benchmarks with AVS-
Rich> 
Rich> Test 1: Teapot. Invoke avs with "avs -timer".  
Rich> Bring up the geometry viewer.
Rich> Load the "teapot.geom" object from the /usr/avs/data/geometry directory.  
Rich> Start it spinning.  Stop it.
Rich> In the window you invoked AVS, you will see the number of
Rich> frames per second (FPS) that your machine produced, and gives
Rich> you a solid number on 3D graphics performance to compare with 
Rich> other machines running AVS.
Rich> 
Rich> Test 2: Mandrill on Teapot.  In the geometry viewer, click on 
Rich> "edit texture", then select the "mandrill.x" image from the 
Rich> /usr/avs/data/image directory.  Click on "set dynamic texture".
Rich> (on machines without hardware texture mapping, you will need to 
Rich> change the "camera" setting from "hardware renderer" to "software renderer".
Rich> Start it spinning. Stop it.  The FPS this time gives you a solid number
Rich> of how well your machine does on 3D graphics performance with 
Rich> texture mapping.
Rich> 
Rich> 
Rich> Here's what I found:
Rich> 
Rich> Tested machines to date (ranked in order of 3D performance):
Rich> 
Rich>                              3D Performance   Texture-Mapped 3D Performance
Rich>                               Teapot (FPS)     Mandrill on Teapot (FPS)  
Rich> Kubota 3400-E210                  72                 10.4
Rich> SGI Crimson VGXT, R4000           60                 15.0
Rich> SGI Indigo2-Extreme,R4000         52                  1.7
Rich> SGI Indigo ELAN, R4000            30                  1.9
Rich> Dec 5000/200 PXG/TURBO            18                  0.5 
Rich> SGI Indigo XS-24, R3000           12                  0.8

Heres what I obtained when performing the aforementioned
benchmarks on my IPX with GX card and 28 Mb memory...

                                        0.9                0.4

I'm not sure what you meant about changing the camera setting to
software renderer, which is what I thought I would need to do,
but it seemed to work anyway.

Pretty slow, huh?

Lloyd



From geir@nilu.no (Geir O. Braathen)
Subject: Movie from raster images
Message-ID: <1994Feb3.145613.17559@nilu.no>
Organization: NILU - Norwegian Institute for Air Research
Date: Thu, 3 Feb 1994 14:56:13 GMT
Lines: 42

Hello,

I am looking for two tools:

1. A utility to dump a part of the X windows screen (a window with graphics
   to be more specific). xwd doesn't meet my needs since I want to make
   many such dumps from a program rather than clicking on the window as
   required by xwd. A variety of xwd which takes as input the window
   coordinates and size would probably be ideal.


2. A tool for playing back the raster files on the screen in rapid succession
   in order to make a movie.


I use the Uniras AGL/FGL Fortran Libraries to produce colour maps on the
screen. I want to visualize the temporal development of various parameters
such as atmospheric temperatures, trace gas concentration fields etc.
We have programs to make plots of one field at the time. It would be
very useful to look at how these fields change with time by playing a large
number (50-100 )of such maps back on the screen as a movie.

The C version of Uniras (AGX Toolmaster) has an animation capability, but
it is unable to store the raster images for later use. Since each picture
takes some time to plot (typically 15-30 seconds on my Sun Sparcstation 2) it
would be nice to store the raster images for playback at a later time.

It would also be nice to be able to compress the raster images to some compact
format, such as jpeg. Is there an X-based movietool around that could display
compressed image files, and is there software to compress Sun raster
images to jpeg or some other compact format?


Best regards

Geir Braathen
Research Scientist
Norwegian Institute for Air Research
P.O. Box 64
N-2001 Lillestrom
Norway
E-mail: geir@nilu.no  


From ferguson@craycos.com (Scott Ferguson)
Newsgroups: comp.graphics.avs
Subject: Re: AVS benchmarking
Date: 3 Feb 1994 08:51:02 -0700
Organization: Cray Computer Corporation
Lines: 33
Message-ID: <2ir6h6$vd@conq2.craycos.com>
References: <CKMKts.M11@silvaco.com>
NNTP-Posting-Host: conq2.craycos.com

We should remember when benchmarking the software renderer that the size
of the window has an AWFUL LOT to do with the number of frames per second.
I assume there's a similar effect with the accelerators, but it's just
less noticeable.

On my Sun4 with an 8-bit non-accelerated graphics card, I got 0.9 frames/sec
on the non-textured teapot when the geometry viewer window was about 512x512,
the default size. When I shrank the window to about half the size, I got
1.6 frames/sec.

Also, if you're running on a remote machine (like a Cray-3), the pictures get
sent by XPutImage across the network. The limits in X11R5 of request buffer
sizes say that if you send a 511x512 image, it has to be broken up into two
separate PutSubImage requests, which slows down the operation a lot, so keeping
your graphics window just smaller than 511x512 helps out a lot too.

On my Cray-2 (being shared by scientists from NCAR and other software 
developers so the machine is severely loaded down) I get 0.41 frames per second
on the non-textured teapot, and if I shrink the window I can triple the speed
because of the reduced ethernet overhead as well as the fewer pixels to fill 
per polygon. When I'm finished debugging the port to our OS, I'll compile AVS5
on the Cray-3, which is approximately twice as fast.

It'd be more ideal running remotely to a supercomputer if you have local
3D graphics power with Z-buffering, etc, but when all your eggs are in the
supercomputer basket you can still get some decent graphics if you keep the
window a little smaller.

-- 
------------------------------------------------------------------------------
Scott Ferguson                                      Cray Computer Corporation
ferguson@craycos.com                                Colorado Springs, CO 80906
------------------------------------------------------------------------------


From ralf@kristall.erdw.ethz.ch (Ralf Grosse Kunstleve)
Newsgroups: comp.graphics.avs
Subject: Q: non-orthogonal transformation of Geometry Data
Date: 03 Feb 1994 19:55:42 GMT
Organization: ETH Zurich - Institute of Crystallography
Lines: 20
Distribution: world
Message-ID: <RALF.94Feb3205542@ruby.ethz.ch>
NNTP-Posting-Host: ruby.ethz.ch

Hello,

Is it possible in principle to perform a non-orthogonal transformation
of Geometry Data. Could somebody provide me with some example code?

To be more specific: In crystallography one often works with a non-orthogonal
basis. In AVS there is the obvious possibility to convert your input data to
the irregular field type and to do all further manipulations on those irregular
fields (e.g. PD module "non_ortho"). Unfortunately some AVS modules can't
handle irregular fields, e.g. arbitrary_slicer.
Now the idea is to do all manipulations on regular fields (since they *are*
regular, just not on a cartesian basis) and put a module just before the
Geometry Viewer, which does the transformation from crystallographic to
cartesian space.

Thank you!

Ralf Grosse Kunstleve
Institute of Crystallography, ETH Zurich, CH
ralf@kristall.erdw.ethz.ch


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - Get_Terrain_Data
Message-ID: <1994Feb9.220706.27769@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:07:06 GMT

Name        : Get_Terrain_DataVersion      : 1.000     Mod Number : 1780
Author      : Jeff Wang, MCNC, Environmetal Programs Group
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/data_input/Get_Terrain_Data
Ported to   : Sun HP
Description : This module allows users to read in sorted ASCII terrain
              data and generate an 2-D irregular field. Together with the
              "Draw Map" module on the ftp site, users can create a 2-D
              terrain plot over the political boundary map. Sample data
              for this module can be found via anonymous ftp to
              avs.ncsc.org, in
              /sample_data/avs_data/Get_Terrain_Data/dma.dat.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - switch_2
Message-ID: <1994Feb9.220759.27838@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:07:59 GMT

Name        : switch_2        Version      : 1.000     Mod Number : 1779
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/switch_2
Ported to   : Sun HP Kubota
Description : This module was submitted as one of a two-module suite - 1)
              switchfire, which switches between two fields, and
              controls flow to downstream modules, and 2) switch, which
              switches between two fields to downstream modules. Switch
              will be useful in keeping the flow of data controllable to
              bottleneck points without requiring the user to
              repeatedly connect and disconnect modules to control
              their source field. The module will send data downstream
              whenever the chosen input is updated, or the A/B switch is
              switched. Both inputs are not required for the module to
              execute. This module allows the user to control the flow of
              AVS fields to downstream modules. Often it may be
              desireable to switch the input path between modules,
              macros or whole branches of a network frequently without
              having to connect and disconnect inputs repeatedly. This
              is especially true in larger networks wherein it would be
              impractical to duplicate the downstream module or tree,
              and switching inputs may require stopping the flow
              executive to allow for time to make the change.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - switchfire
Message-ID: <1994Feb9.220831.27909@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:08:31 GMT

Name        : switchfire      Version      : 1.000     Mod Number : 1778
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/switchfire
Ported to   : Sun HP Kubota
Description : This module was submitted as one of a two-module suite - 1)
              switchfire, which switches between two fields, and
              controls flow to downstream modules, and 2) switch, which
              switches between two fields to downstream modules.
              Switchfire will be useful in keeping the flow of data
              controllable to bottleneck points without requiring the
              user to repeatedly connect and disconnect modules to
              control their source field and execution. The module will
              switch between two input fields and send data downstream.
              In control mode, data is sent when triggered by the "send
              data" button, or, in continuous mode, whenever the chosen
              input is updated, the A/B choice is switched, or the send
              data button is triggered. Both inputs are not required for
              the module to execute. This module allows the user to
              control the flow of AVS fields to downstream modules. Often
              it may be desireable to switch the input path between
              modules, macros or whole branches of a network frequently
              without having to connect and disconnect inputs
              repeatedly. Similarly, it is often desireable to stop,
              start, or trigger data flow and execution to individual
              modules or branches of a network. This is especially a
              problem with modules that overwrite existing files
              whenever they are updated, or are time consuming to run when
              frequently not necessary.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - real_control
Message-ID: <1994Feb9.220900.27978@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:09:00 GMT

Name        : real_control    Version      : 1.000     Mod Number : 1777
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/real_control
Ported to   : Sun HP Kubota
Description : This module was submitted as one of a four-module suite - 1)
              dual control - to control flow of field and integers to
              downstream modules, 2) field control - to control flow of an
              AVS image to downstream modules, 3) integer control - to
              control flow of an integer to downstream modules, and 4)
              real control - to control flow of an real to downstream
              modules. This module serves the same purpose as Integer
              Control except for floats instead of integers. Its
              operation is identical. See Field Control for further
              info. An example of its use is to allow animate float to
              remain connected to a number of float ports on one or several
              modules while allowing the user to select which of these
              downstream modules will be effected. Provision is also
              made for manually sending single real updates downstream.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - integer_control
Message-ID: <1994Feb9.221213.28127@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:12:13 GMT

Name        : integer_control Version      : 1.000     Mod Number : 1776
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/integer_control
Ported to   : Sun HP
Description : This module was submitted as one of a four-module suite - 1)
              dual control - to control flow of field and integers to
              downstream modules, 2) field control - to control flow of an
              AVS image to downstream modules, 3) integer control - to
              control flow of an integer to downstream modules, and 4)
              real control - to control flow of an real to downstream
              modules. This module serves the same purpose as Field
              Control except for integers instead of fields. Its
              operation is identical. See Field Control for further
              info. An example of its use is to allow animate integer to
              remain connected to a number of integer ports on one or
              several modules while allowing the user to select which of
              these downstream modules will be effected. Provision is
              also made for manually sending single integer updates
              downstream.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - field_control
Message-ID: <1994Feb9.221240.28198@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:12:40 GMT

Name        : field_control   Version      : 1.000     Mod Number : 1775
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/field_control
Ported to   : Sun HP Kubota
Description : This module was submitted as one of a four-module suite - 1)
              dual control - to control flow of field and integers to
              downstream modules, 2) field control - to control flow of an
              AVS image to downstream modules, 3) integer control - to
              control flow of an integer to downstream modules, and 4)
              real control - to control flow of an real to downstream
              modules. Field control will be useful in keeping the flow of
              data controllable to bottleneck points without requiring
              the user to repeatedly connect and disconnect modules to
              control their execution. This module allows the user to
              control the flow of an AVS field to downstream modules.
              Often there will be modules, (such as tracer), macros or
              whole branches of a network that take a long time to execute
              and may be undesireable in any given run of data through the
              network. Without this module, the only way to activate or
              deactivate such a module or branch would be to connect or
              disconnect it's input, suspend the flow executive of the
              whole network, or to deactivate it through its editor
              dialog. Another example of where this module will be useful
              is with modules that overwrite files everytime their input
              is updated. Field control can be used to only triger such a
              module as needed.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - dual_control
Message-ID: <1994Feb9.221303.28264@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:13:03 GMT

Name        : dual_control    Version      : 1.000     Mod Number : 1774
Author      : A. Jacobson, UCLA Dept. Pharmacology / Crump Institute
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/dual_control
Ported to   : Sun HP Kubota
Description : This module was submitted as one of a four-module suite - 1)
              dual control - to control flow of field and integers to
              downstream modules, 2) field control - to control flow of an
              AVS image to downstream modules, 3) integer control - to
              control flow of an integer to downstream modules, and 4)
              real control - to control flow of an real to downstream
              modules. Dual control will be useful in keeping the flow of
              data controlable to modules that take integers and fields
              and fire when either input port (or parameter) is changed.
              This allows the user the flexibility to update either of the
              module inputs without the module firing. The only other way
              to do this with continuing changes would be to connect and
              disconnect either or both input ports or stop the network
              flow executive. This module is a combination of the field
              control and integer control modules. See the
              documentation for these modules for further details. An
              example of where this module will be useful is with modules
              that overwrite files everytime their input is updated.
              Dual control can be used to only triger such a module as
              needed.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Updated module at IAC - write_A60_yuv
Message-ID: <1994Feb9.221325.28330@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:13:25 GMT

Name        : write_A60_yu    Version      : 3.000     Mod Number : 1106
Author      : Terry Myerson, International AVS Center (NCSC)
Submitted   : 11/27/91        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/data_output/write_A60_yu
Ported to   : Sun HP IBM
Description : Intended to be used with the AVS Animator application. The
              module hooks right up to the Animator module or the
              read_frame_seq module. This module accepts images, one at
              a time. The image is then converted into the Abekas A60 yuv
              format - with the same dimensions of the Abekas A60 yuv
              format - and output to a file in the Frame directory with a
              name corresponding to the frame number. ( i.e. Frame Number
              20 -- Filename - 20 ) If the image is not the same size as the
              A60 image standard, then the AVS image is centered within the
              view of the Abekas. This works well when you consider the
              degradation of the image borders on its path to NCSC. This
              module was adapted from Ian Currington's write_abekas
              module and Abekas Inc.'s sample to_yuv program in their
              manual. It writes out a temporary file in between
              conversions. The module works fine - but it isn't pretty.
              Version 3 update provided by the UCLA School of Medicine's
              Andy Jacobson. The update includes bug fixes, and features
              wherein the user can name the file as with the
              animate_filename module, but with a fixed .yuv siffix. The
              user also has the choice of compressing the (huge!) output
              files. In some nets it may be desireable to allow or prevent
              overwriting of files, so a switch is now provided for this.
              Also you can supply your favorite compression program to
              this module, if desired.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - image_to_YUV
Message-ID: <1994Feb9.221358.28396@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:13:58 GMT

Name        : image_to_YUV    Version      : 1.000     Mod Number : 1773
Author      : P. Fulconis, CICG computer center of Grenoble University,
              France
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/data_output/image_to_YUV
Ported to   : IBM
Description : This is one of two modules written to translate AVS
              animations into mpeg format - image_to_YUV and
              image_to_PPM. Both formats can be read by mpeg encoder. A
              ppm image will have a memory size twice bigger than the yuv
              one, but it can be visualised by xv. You will need to make use
              of mpeg_encode and mpeg_play, both public domain
              utilities available via anonymous ftp from the University
              of California at Berkeley. These modules also make use of
              the San Diego Supercomputer Center's Image Tools, also
              available via anonymous ftp. NOTE TO READER - IF YOU KNOW P.
              FULCONIS OR HIS EMAIL ADDRESS, PLEASE SEND HIS/HER CONTACT
              INFO TO THE IAC AT avs@ncsc.org - WE HAVE A QUESTION FOR
              HIM/HER ABOUT ONE OF THE LIBRARIES - THANK YOU

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - image_to_PPM
Message-ID: <1994Feb9.221425.28466@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:14:25 GMT

Name        : image_to_PPM    Version      : 1.000     Mod Number : 1772
Author      : P. Fulconis, CICG computer center of Grenoble University,
              France
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/data_output/image_to_PPM
Ported to   : IBM
Description : This is one of two modules written to translate AVS
              animations into mpeg format - image_to_YUV and
              image_to_PPM. Both formats can be read by mpeg encoder. A
              ppm image will have a memory size twice bigger than the yuv
              one, but it can be visualised by xv. You will need to make use
              of mpeg_encode and mpeg_play, both public domain
              utilities available via anonymous ftp from the University
              of California at Berkeley. These modules also make use of
              the San Diego Supercomputer Center's Image Tools, also
              available via anonymous ftp. NOTE TO READER - IF YOU KNOW P.
              FULCONIS OR HIS EMAIL ADDRESS, PLEASE SEND HIS/HER CONTACT
              INFO TO THE IAC AT avs@ncsc.org - WE HAVE A QUESTION FOR HIM
              ABOUT ONE OF THE LIBRARIES - THANK YOU

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Updated module at IAC - new_arbitrar
Message-ID: <1994Feb9.221454.28543@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:14:54 GMT

Name        : new_arbitrar    Version      : 2.000     Mod Number : 1003
Author      : Wes Bethel, Lawrence Berkeley Laboratory
Submitted   : 10/10/91        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/mappers/new_arbitrar
Ported to   : Sun HP IBM Kubota
Description : The new arbitrary slicer module extracts a 2d slice from a 3d
              field. Output consists of a 2d field (as opposed to
              geometry, as in the "normal" arbitrary slice module) which
              may be piped to other modules which process 2d fields. The
              position, size and resolution of the slice plane are under
              user control via the parameters. For all input fields, the
              module will determine "sane" parameter settings for
              extracting initial slice. The latest update provides
              PRELIMINARY support for irregular fields. This HAS ITS
              LIMITATIONS - it uses a brute-force algorithm WHICH IS VERY
              SLOW; as space subdivision on warped grids is still a
              research topic.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - integer_step
Message-ID: <1994Feb9.221605.28644@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:16:05 GMT

Name        : integer_step    Version      : 1.000     Mod Number : 1771
Author      : Young S. Ham, Naval Research Laboratory
Submitted   : 02/09/94        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/data_input/integer_step
Ported to   : Sun HP Kubota
Description : The integer_step module supplies an integer value to one or
              more receiving modules from an intial integer. The current
              integer is incremented by step_size by means of a mouse
              click on a button. This module is very useful when viewing
              images or to fire an input value in order is desired. One does
              not have to keep typing an integer or play with the dial
              widget for a new input value.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - Field_Math
Message-ID: <1994Feb9.221650.28714@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:16:50 GMT

Name        : Field_Math      Version      : 1.000     Mod Number : 1770
Author      : Krzysztof (Chris) S. Nowinski, Institute of Applied
              Mathematics and Mechanics, Warsaw University
Submitted   : 02/03/94        Last Updated : 02/03/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/Field_Math
Ported to   : Sun HP
Description : The module does some field mathematics -- I am intended to
              talk about it at the AVS94. The module interprets a short (up
              to 60 lines) program written in AVSFOL (A Very Simple Field
              Operation Language). AVSSPL admits two types of variables
              - simple variables with values being single float numbers;
              field variables with values being float arrays of
              dimensions determined by the output field.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Updated module at IAC - city_scape
Message-ID: <1994Feb9.221718.28784@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:17:18 GMT

Name        : city_scape      Version      : 3.000     Mod Number : 1185
Author      : Ian Curington, Advanced Visual Systems
Submitted   : 05/04/92        Last Updated : 02/09/94  Language   : C
Module path : avs.ncsc.org:avs_modules/mappers/city_scape
Ported to   : Sun HP IBM Kubota
Description : The city scape module converts a two-dimensional floating
              point field into group of 3D blocks, represented as a
              GEOM-format mesh. Each element of the field is mapped to a
              "building". The height of each block above each point is
              proportional to the scalar value of the field. This is my
              attempt at "Business Graphics", for a multi-dimensional
              telecommunications statistics data application, except
              in interactive 3D. Version 2 is a substantial upgrage from
              version 1 at the IAC. This module provides 3D block diagram,
              with averaged statistics on side shadow panels, and
              annotation. A transparent theshold sheet is superimposed
              on city. Blocks can be picked for interogation of cell
              number. parts may be turned on or off for clarity. Two scalar
              fields can be displayed, one contolling height, the other
              color, with two independent side stacked side bar
              statistics. Version 3 update draws the top face of picked
              block in white hi-lite.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Anyone know P. Fulconis of CICG in France?
Message-ID: <1994Feb9.222224.28909@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Wed, 9 Feb 1994 22:22:24 GMT

Does anyone know P. Fulconis of the CICG computer center 
of Grenoble University, France?

If you have an email address, we'd greatly appreciate it
if you forwarded it to us at avs@ncsc.org.  Thanks!

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From erich@ee.pdx.edu (Erich S. Boleyn)
Newsgroups: comp.graphics.avs
Subject: Re: AVS benchmarking
Date: 3 Feb 1994 20:59:55 -0800
Lines: 42
Message-ID: <2iskob$qsc@cruella.ee.pdx.edu>
References: <CKMKts.M11@silvaco.com>
NNTP-Posting-Host: cruella.ee.pdx.edu
X-Newsreader: NN version 6.4.19 #2

lloyd@silvaco.com (Lloyd Evans) writes:


>Rich> Graphics Benchmarks with AVS-
>Rich> 
>Rich> Test 1: Teapot. Invoke avs with "avs -timer".  

...[teapot test]...

...[teapot with mandrill test]...

>Rich> Here's what I found:
>Rich> 
>Rich> Tested machines to date (ranked in order of 3D performance):
>Rich> 
>Rich>                          3D Performance   Texture-Mapped 3D Performance
>Rich>                           Teapot (FPS)     Mandrill on Teapot (FPS)  
>Rich> Kubota 3400-E210                72                 10.4
>Rich> SGI Crimson VGXT, R4000         60                 15.0
>Rich> SGI Indigo2-Extreme,R4000       52                  1.7
>Rich> SGI Indigo ELAN, R4000          30                  1.9
>Rich> Dec 5000/200 PXG/TURBO          18                  0.5 
>Rich> SGI Indigo XS-24, R3000         12                  0.8

Here's mine (using AVS 4):

Stardent Vistra 800EX                  11.3                6.0

Sort of a funny ratio, I must say.  Oh, yes, that's using gouraud shading,
it turns out to be only slightly faster than even unlighted polygons,
which was about 12.5 fps.  Using "points" gave me about 15.5 fps.

The more complex the scene, the better (relatively) this machine
performs.

Erich Boleyn

-- 
             "I haven't lost my mind; I know exactly where it is."
   / --  Erich Stefan Boleyn  -- \        --=> *Mad Genius wanna-be* <=--
  { Honorary Grad. Student (Math) }-->  Internet E-mail: <erich@ee.pdx.edu>
   \  Portland State University  /       WARNING: INTERESTED AND EXCITABLE


From zaxin@bigwpi.WPI.EDU (Benjamin C. Lee)
Newsgroups: comp.graphics.avs
Subject: Use of object utility routines
Date: 5 Feb 1994 21:09:21 GMT
Organization: Worcester Polytechnic Institute
Lines: 15
Message-ID: <2j11u1$1i1@bigboote.WPI.EDU>
NNTP-Posting-Host: bigwpi.wpi.edu


There are several routines listed in the AVS developer's guide which
take in a GEOMobj type as input (i.e. GEOMset_color ) In every example
that I've seen the object itself was first created and then sent to
the procedure.  Is it possible to select an already existing GEOMobj
from an edit list and use that as input to the GEOMset_color.  If so,
how?



-- 
--
Disclaimer: I don't even know her!  | Benjamin "It's the hole I've dug" Lee
Nothing in Moderation; Do what you do... Do it well  |  zaxin@wpi.wpi.edu
GCS -d+(---) p c++++ l u++ e+/* m--- s+++/+ n- h* f+(*) g++ w+++ t--- r+ y*


From ok@wsvst31.site.uni-wuppertal.de (Olaf H. Kelle)
Newsgroups: comp.graphics.avs
Subject: Wanted: chart visualisation module ?
Date: Mon, 7 Feb 1994 13:42:09
Organization: Universitiy of Wuppertal FB 14
Lines: 16
Distribution: world
Message-ID: <ok.28.000DB44D@wsvst31.site.uni-wuppertal.de>
NNTP-Posting-Host: wsvst32.site.uni-wuppertal.de
X-Newsreader: Trumpet for Windows [Version 1.0 Rev A]

For my doctoral thesis I have to generate some simple 3D-bar charts. But these 
should have a professional touch. So the module should have the opportunity to 
generate shadows and transparency. The integration of a background texture 
would be nice too.

Is something like this available ?

Thank you very much for your help.

Olaf

Dipl.-Ing. Olaf H. Kelle
University of Wuppertal, Germany
Dep. for Transportation Safety and Simulation
EMail: olaf@wsvst31.site.uni-wuppertal.de
          kelle@wrcd1.urz.uni-wuppertal.de


From hansen@acl.lanl.gov (Charles D. Hansen)
Subject: Re: AVS benchmarking (How does YOUR system rate?)
In-Reply-To: rsignell@crusty.er.usgs.gov's message of Wed, 2 Feb 1994 19:58:45 GMT
Message-ID: <HANSEN.94Feb7122109@gummy.acl.lanl.gov>
Sender: news@newshost.lanl.gov
Reply-To: hansen@acl.lanl.gov
Organization: Advanced Computing Lab, LANL, NM
References: <CKM4ty.Kwr@netnews.whoi.edu>
Date: Mon, 7 Feb 1994 19:21:09 GMT
Lines: 19

In article <CKM4ty.Kwr@netnews.whoi.edu> rsignell@crusty.er.usgs.gov (Richard P. Signell) writes:
   Graphics Benchmarks with AVS-

   Test 1: Teapot. Invoke avs with "avs -timer".  

   Test 2: Mandrill on Teapot.  In the geometry viewer, click on 


These benchmarks are only indicative of real application if you creat
*small* geometries (such as the teapot).  We have found that much
larger geometries are the norm ... at least for us.  Thus, this
benchmark would have very little bearing on real applications.  A much
better benchmark would be to read in a large geometry (the crib for
example) and render that.  You'll a see much greater difference due to
the traversal of the geometry structure.  At least that's our
experience .... you mileage may vary.

Chuck Hansen
hansen@acl.lanl.gov


From dxr103@cac.psu.edu (Tad Rollow)
Newsgroups: comp.graphics.avs
Subject: Re: AVS benchmarking
Date: 8 Feb 1994 17:35:13 GMT
Organization: Penn State University
Lines: 24
Message-ID: <2j8igh$1m9@genesis.ait.psu.edu>
References: <CKMKts.M11@silvaco.com> <2iskob$qsc@cruella.ee.pdx.edu>
NNTP-Posting-Host: curtiss.cac.psu.edu

In article <2iskob$qsc@cruella.ee.pdx.edu>,
Erich S. Boleyn <erich@ee.pdx.edu> wrote:
> Graphics Benchmarks with AVS-

This seems to be hitting a related theme to something I've been curious
about:

We have an HP 720 with a CRX 24 graphics card, and we're stupidly NOT
using the PHIGS run time library!  We'll order it, but does anyone
out there know what performance advantage this'll bring us?

And similarly, for some couple of thousand dollars, we can get Z buffering
in the form of a CRX 24 Z upgrade.  What do we see then?
I think the HP is potentially a much faster machine than what we see now.

Please excuse my inquiries if they seem merely labor-avoiding; I am also
going to ask the friendly folks at AVS whether they have this stuff
documented, and perhaps HP can attempt to sell me hardware if they have
it documented too.
But I thought someone out there might have direct experience.

Thanks
Tad Rollow   WB3KUZ   tad@sabine.arl.psu.edu
Computational Acoustics Lab, Penn State Acoustics Dept.


From hstroyan@fc.hp.com (Howard Stroyan)
Subject: Re: AVS benchmarking
Sender: news@fc.hp.com (news daemon)
Message-ID: <CKxEF1.7sr@fc.hp.com>
Date: Tue, 8 Feb 1994 21:59:25 GMT
References: <CKMKts.M11@silvaco.com> <2iskob$qsc@cruella.ee.pdx.edu> <2j8igh$1m9@genesis.ait.psu.edu>
Nntp-Posting-Host: hphonk.fc.hp.com
Organization: Hewlett-Packard Fort Collins Site
X-Newsreader: TIN [version 1.2 PL1.4]
Lines: 31

Tad Rollow (dxr103@cac.psu.edu) wrote:
: In article <2iskob$qsc@cruella.ee.pdx.edu>,
: Erich S. Boleyn <erich@ee.pdx.edu> wrote:
: > Graphics Benchmarks with AVS-

: This seems to be hitting a related theme to something I've been curious
: about:

: We have an HP 720 with a CRX 24 graphics card, and we're stupidly NOT
: using the PHIGS run time library!  We'll order it, but does anyone
: out there know what performance advantage this'll bring us?

: And similarly, for some couple of thousand dollars, we can get Z buffering
: in the form of a CRX 24 Z upgrade.  What do we see then?
: I think the HP is potentially a much faster machine than what we see now.

I happen to have a HP720 with a CRX24Z and AVS 5.0 in front of me.
If I perform the little spinning teapot experiment I get the following
results:

	AVS software renderer                 -  2.001290 FPS
	CRX24 (Z disabled) Hardware renderer  -  5.179334 FPS
	CRX24Z Hardware renderer              - 23.776882 FPS
	
As noted in another posting, this is a somewhat simplistic measure,
but it give some rough idea of relative performance.

--
Howard Stroyan
Hewlett-Packard
Advanced Systems Division / Graphics S/W Lab (ASD/GSL)     hstroyan@fc.hp.com


From juanman@picasso.ucsd.edu (Juan Manuel Castaneda)
Newsgroups: comp.graphics.avs
Subject: Map Outlines in AVS
Date: 9 Feb 1994 01:24:28 GMT
Organization: A poorly-installed InterNetNews site
Lines: 17
Message-ID: <2j9e0c$igs@network.ucsd.edu>
NNTP-Posting-Host: picasso.ucsd.edu
X-Newsreader: TIN [version 1.2 PL2]

Does a module or technique exist to create/import geographical map outlines
in AVS?  The ideal situation would be to take map and orography data 
from say, NCAR Graphics, and save it as a geometry.

Thanks, in advance, for any info.

--Juan
--
***************************************************************************
Juan Manuel Castaneda Valdez 		Climate Research Division
juanman@picasso.ucsd.edu		Scripps Institution of Oceanography
Voice:	(619)534-2241			University of California, San Diego
Fax:	(619)534-8561			La Jolla, CA 92093-0224

"What does a fish know about the water in which he swims?" A. Einstein
Spanish Translation:  "Camarron que se duerme, se lo lleva la corriente."
*************************************************************************** 


From larryg@avs.com (Larry Gelberg)
Newsgroups: comp.graphics.avs
Subject: Re: References
Date: 9 Feb 1994 15:12:59 GMT
Organization: Advanced Visual Systems Inc.
Lines: 29
Distribution: world
Message-ID: <2jauhr$k5r@nda.nda.com>
References: <1994Feb8.203656.163165@embl-heidelberg.de>
NNTP-Posting-Host: phobos.avs.com
X-Newsreader: TIN [version 1.2 PL1]

Stefan Herr (herr@EMBL-Heidelberg.DE) wrote:

: is there a list of publications concerning the use of AVS in different fields as
: well as a general description of AVS - perferable in international journals.

: Please reply directly to herr@embl-heidelberg.de

: Thansk in advance, stefan

Stefan,

The Proceedings from the last two AVS User Group meetings contain
dozens of papers broken down by subject (Medical Imaging, Chemistry,
Image Analysis, etc)  Generally, these are application notes: "This is
how we use AVS to do ....".  These are (IMHO) an invaluable resource
for learning who your colleagues are and seeing their approach to using
AVS to solve their problems. 

Contact avs@ncsc.org for information on ordering back issues of these
Proceedings and please consider coming to the AVS '94 Users Group 
Meeting being held here in Boston, May 2-4, 1994.

larryg

--
=== Larry Gelberg ============================ larryg@avs.com =======
      Advanced Visual Systems Inc. (AVS Inc.)
      300 Fifth Ave, Waltham, MA 02154
===== Tel: 617-890-4300 = Fax: 617-890-8287 =========================


From wang@willis.cis.uab.edu (Yun Wang)
Subject: AVS module
Message-ID: <1994Feb9.214117.16709@cis.uab.edu>
Sender: wang@cis.uab.edu (Yun Wang)
Organization: University of Alabama at Birmingham
Date: Wed, 9 Feb 1994 21:41:17 GMT
Lines: 21


Does anybody out there know what kind of algorithm is the module isosurface
used in AVS? Is it marchingcube?

Also, can anybody tell me how to control the threshold panel of isosurface
module in a better precision? It seems that it jumps from 209 something 
directly to 212, and that is very annoying!

Finally, I have some trouble converting volume to field data in AVS. IS
there any separation character I need to separate the header and the
binary area? Do I just simply print the header and copy the binary area
from volume format to field data?

--Yun Wang
wang@cis.uab.edu

-- 
-------------------------------
Yun Wang
Research Assistant
Dept. of Computer and Information Science


From tpowell@dsd.es.com (Tim Powell)
Newsgroups: comp.graphics.avs
Subject: AVS Benchmarking: E&S Freedom/AVS 5.01
Date: 9 Feb 1994 23:03:41 GMT
Organization: Evans & Sutherland Computer Corp., Salt Lake City, UT
Lines: 98
Distribution: world
Message-ID: <2jbq4d$1hf@orca.es.com>
Reply-To: tpowell@dsd.es.com (Tim Powell)
NNTP-Posting-Host: 130.187.90.18

To: rsignell@crusty.er.usgs.gov (Richard P. Signell)
Subject: Re: AVS benchmarking (How does YOUR system rate?)
 
In article <CKM4ty.Kwr@netnews.whoi.edu> you write:
 
> Tested machines to date (ranked in order of 3D performance):
> 
>                              3D Performance   Texture-Mapped 3D Performance
>                               Teapot (FPS)     Mandrill on Teapot (FPS)  
> Kubota 3400-E210                  72                 10.4
> SGI Crimson VGXT, R4000           60                 15.0
> SGI Indigo2-Extreme,R4000         52                  1.7
> SGI Indigo ELAN, R4000            30                  1.9
> Dec 5000/200 PXG/TURBO            18                  0.5 
> SGI Indigo XS-24, R3000           12                  0.8
> 
> 
> Please help me add to this list!

Here are the numbers for a Sun SPARCStation 10/41 with 
E&S Freedom Series 3200 graphics:

Teapot (FPS): 88
Texture-Mapped Mandrill on Teapot (FPS): 77

If vertical retrace were turned on, both numbers would drop to 72,
the refresh rate of the display. You've probably had it pointed out
to you that since some of the machines on the list are vertical
retrace limited, the teapot geometry has too few polygons (3936)
to give a good indication of their peak performance. The carib 
geometry, however, has 26,529 tri-strip polygons, is gouraud-shaded,
and has color specified per vertex. Benchmarking with it gives
a better indication of peak graphics performance; here are the
numbers for the Sun/E&S configuration stated above:

Carib (FPS): 32
Texture-Mapped Mandrill on Carib (FPS): 22 
				(=583,638 textured polys/sec!)

It would be interesting for you (or others) to post the carib 
numbers for the other systems mentioned above. 

Another consideration while doing texture benchmarking is the type 
of texture-mapping being done. Following your procedures, I believe 
the default in AVS is point-sampled texture, which is the poorest 
quality of texture mapping possible. Mipmapping and blending can 
be done to improve the quality of texture-mapping at some cost 
to performance. 

The numbers quoted for the Sun/E&S system above are not the 
fastest possible. A SPARCStation 10/51 would provide more CPU
horsepower to feed the graphics accelerator, and the E&S
Freedom Series 3300 provides 50% better performance than the
3200 (actual mileage may vary...). I don't have a system with 
this configuration available to me right now, but I'll post 
the numbers when I do.

Here's some information on the version of AVS running on
the Evans & Sutherland hardware:

AVS 5.01 FOR EVANS & SUTHERLAND 
FREEDOM SERIES GRAPHICS ON SUN SPARCSTATIONS

A new version of AVS 5.01 has been released by AVS last month
for the E&S Freedom Series hardware with support for the
following:

o 2D Texture Mapping with the following features:
  -indexed texture support
  -ability to translate and scale textures
  -texture tiling
  -alpha texturing (texture transparency)
  -automatic interpolation of texture sizes
  -texture filtering
  -trilinear mipmap support
o Transparency
o Antialiased lines and markers with your choice of alpha
  blending or shadowfax blending
o Edge color specification (for outline gouraud)
o Dual-headed support, so that the user interface can
  run on one head and the graphics can run on the other
o Abitrary clipping support

For more information about this version of AVS, call AVS 
at (800) 428-7001.

For more information about the E&S Freedom Series, call E&S 
at (800) 367-7460.

I know this is overkill, but I hope this helps!

Tim 

Tim Powell                        | email: tpowell@dsd.es.com
ISV Relations Specialist          | Phone: (801) 582-5847
Evans & Sutherland                |   Fax: (801) 582-0524
Design Systems Division           |   ftp: ftp.es.com (130.187.2.100)
---------------------------------------------------------------------


From wjj@burford.whoi.edu (W.J. Jenkins)
Subject: Re: AVS module
Message-ID: <CL0KIy.J6n@netnews.whoi.edu>
Sender: news@netnews.whoi.edu
Nntp-Posting-Host: burford
Organization: whoi
References:  <1994Feb9.214117.16709@cis.uab.edu>
Date: Thu, 10 Feb 1994 15:04:10 GMT
Lines: 13

In article <1994Feb9.214117.16709@cis.uab.edu>, wang@willis.cis.uab.edu (Yun Wang) writes:
|> 
|> Also, can anybody tell me how to control the threshold panel of isosurface
|> module in a better precision? It seems that it jumps from 209 something 
|> directly to 212, and that is very annoying!
|> 


If you click on the center of the dial, a window will pop up to allow you
to modify the values by typing them in.  You can also change the ranges.
Give it a try.

bill


From avs@doppler.NCSC.org (AVS account)
Subject: 7 New AVS/IDL Mods at IAC
Message-ID: <1994Feb17.212529.7053@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: avs@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 17 Feb 1994 21:25:29 GMT

Module paths: avs.ncsc.org:avs_modules/filters/<module_name>
                 which are just 7 different links to
              avs.ncsc.org:avs_modules/filters/IDL_modules
                 where all the AVS IDL modules are stored together.
                 The links are set up so the IAC database and catalog
                 scripts will find the individual modules correctly.

Author      : Joshua Goldstein, Research Systems Incorporated
Submitted   : 02/15/94        Last Updated : 02/15/94  Language   : C
Ported to   : Sun IBM
Description : This is a suite of seven AVS/IDL modules. The modules
              are all in the FILTERS subdirectory at the IAC. They are IDL -
              a UNIX utility that gets IDL and AVS communicating; and six
              AVS modules that convert data between IDL and AVS -
              field_to_float_parm, field_to_int_parm,
              field_to_string_parm, float_parm_to_field,
              int_parm_to_field, and string_parm_to_field. These
              modules have only been tested on the RS 6000 and Sun
              Sparcstation. Binaries have been compiled on the HP 700 but
              its usability is unknown. Binary executables are stored at
              the IAC for all three of these platforms. There are also some
              demo files at the IAC, stored in the file demo.tar. You must
              be running IDL 3.1 or later in order to use these. No
              guarantee is made that this version of avsidl will work with
              future versions of AVS and/or IDL. In fact, I know that
              avsidl will be completely reworked for for IDL 4.0.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Updated module at IAC - field_to_globe
Message-ID: <1994Feb17.212618.7160@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: avs@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 17 Feb 1994 21:26:18 GMT

Name        : field_to_globe  Version      : 2.000     Mod Number : 1595
Author      : Ed Bender, Convex Computer Corporation
Submitted   : 03/04/93        Last Updated : 02/15/94  Language   : C
Module path : avs.ncsc.org:avs_modules/filters/field_to_globe
Ported to   : Sun HP IBM
Description : This module is used to wrap field data around a globe. It does
              a pretty good job of allowing the user to specify where on the
              globe to put the corners of the input field or to use the
              points data of an input irregular field as the longitude and
              latitude of each point. If the data is 3-D, then concentric
              globes are generated with the radius determined by the
              "radius min" and "radius max" parameters, unless the data
              is 3-space and the "use height data" toggle is on. Version 2
              enhancements were made to this module by Scott Ferguson of
              Cray Computer Corporation, Colorado Springs, CO - Inner
              loops were restructured to allow vectorization. Also, for
              scalar input fields, the scalar value can be used as
              elevation data. And, for 2-vector or 3-vector input
              fields, the vector data can be converted from polar to
              cartesian coordinates as well. For example, a model of wind
              vectors in a global climate model needed to be overlayed on a
              globe. By selecting "Convert Data", the data can later be
              passed into hedgehog to show the arrows wrapped around a
              sphere. A parameter called "Elev Scale" was added to scale
              the elevation data when using a scalar input field as an
              elevation map.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: New module at IAC - 3D_contour_map
Message-ID: <1994Feb17.212647.7257@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: thorpe@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 17 Feb 1994 21:26:47 GMT

Name        : 3D_contour_map  Version      : 1.000     Mod Number : 1781
Author      : Krzysztof (Chris) S. Nowinski, Institute of Applied
              Mathematics and Mechanics, Warsaw University
Submitted   : 02/15/94        Last Updated : 02/15/94  Language   : C
Module path : avs.ncsc.org:avs_modules/mappers/3D_contour_map
Ported to   : Sun HP
Description : The 3D contour map routine creates contour map of a given 2D
              scalar float field together (optionally) with a surface
              equivalent to the surface generated by the field to mesh
              routine. The algorithm used is a subset of the more familiar
              3d marching cubes algorithm. This is a modification of the
              field_to_con module by Wes Bethel.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From avs@doppler.NCSC.org (AVS account)
Subject: Updated module at IAC - field_to_con
Message-ID: <1994Feb17.212723.7363@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: avs@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Thu, 17 Feb 1994 21:27:23 GMT

Name        : field_to_con    Version      : 2.000     Mod Number : 1024
Author      : Wes Bethel, Lawrence Berkeley Laboratory
Submitted   : 10/10/91        Last Updated : 2/15/94   Language   : C
Module path : avs.ncsc.org:avs_modules/mappers/field_to_con
Ported to   : Sun HP IBM
Description : The field to contour module creates colored contour lines
              around regions in which the threshold specified by the
              "level" parameter is crossed. The module can process only
              2d datasets, but these may may exist in 3-space ( if the field
              is irregular) or in 2-space (rectilinear or uniform). The
              algorithm used is a subset of teh more familiar 3d marching
              cubes algorithm. Version 2 update provided by Krzysztof
              (Chris) S. Nowinski of Warsaw University's Institute of
              Applied Mathematics and Mechanics. Enhancements include
              Generating contours as polylines and not disjoint lines
              for faster rendering; Generating whole range of contours
              by a single module; Generating surfaces (these produced by
              field_to_mesh do not fit into dimensions, sometime);
              Optional plane and/or 3D maps; and Second field and
              colormap input to color surfaces.

International AVS Center
North Carolina Supercomputing Center
avs@ncsc.org


From tim@osc.EDU (Tim Rozmajzl)
Newsgroups: comp.graphics.avs
Subject: Projected Vectors
Date: 21 Feb 1994 13:41:31 GMT
Organization: Ohio Supercomputer Center
Lines: 18
Distribution: world
Message-ID: <2kadmb$nlk@mane.cgrg.ohio-state.edu>
NNTP-Posting-Host: killian.osc.edu

Greetings,

I'm searching for a module that will, given a 2-D orthogonal
slice of a 3-D irregular 3-Vector field, draw vector arrows
projected onto the irregular 2-D slice.  Anyone tried it?

Thanks in advance,

  Tim

-- 
o------------------------The Ohio Supercomputer Center ---------------o
o  Tim Rozmajzl                   |  Phone: (614) 292-3105            o
o  User Services		  |  FAX:   (614) 292-7168            o
o  The Ohio Supercomputer Center  |  Email: tim@osc.edu               o
o  1224 Kinnear Rd.               |                                   o
o  Columbus, Ohio 43212           |                                   o
o---------------------------------------------------------------------o


From wang@willis.cis.uab.edu (Yun Wang)
Subject: Geometry to image
Message-ID: <1994Feb21.231019.28830@cis.uab.edu>
Sender: wang@cis.uab.edu (Yun Wang)
Organization: University of Alabama at Birmingham
Date: Mon, 21 Feb 1994 23:10:19 GMT
Lines: 16

I have a question about the gemoetry viewer:

I have modified the geometry of one of my object like changing the color and 
lighting, and normalize etc. I want to save that particular image. Is there
any feature under geometry viewer I can save that particular images? Or do
I have to save it as adding .scr and use Network Editor to read that geometry 
and output to pixmap(or write image)?


Yun Wang
wang@cis.uab.edu
-- 
-------------------------------
Yun Wang
Research Assistant
Dept. of Computer and Information Science


From ianh@resmel.bhp.com.au (Ian Hoyle)
Newsgroups: comp.graphics.avs
Subject: WANTED: tool (or module) to create simple geometry
Date: 21 Feb 94 12:08:45 GMT
Organization: BHP Research - Melbourne Laboratories
Lines: 17
Message-ID: <ianh.761832525@monster>
NNTP-Posting-Host: monster.resmel.bhp.com.au

I need to create the geometry to display a "rubik cube" that has specified
colours for each of the 9 squares visible on each face of the cube.

Any suggestions for quickly whipping this up for display in AVS ??

Cheers,

	ian
-- 
   /\/\     :  Ian Hoyle,  Senior Research Scientist
  / / /\    :  BHP Research - Melbourne Laboratories
 / / /  \   :  245 Wellington Rd, Mulgrave, 3170, AUSTRALIA
/ / / /\ \  :  Phone    +61-3-560-7066
\ \/ / / /  :  E-mail   ianh@resmel.bhp.com.au
 \  / / /   :  
  \/\/\/    : "perl - the swiss army chainsaw of UNIX tools"
            :               -- Rob Kolstad


From keith@earth.ox.ac.uk (Keith Refson)
Subject: Re: ANSI AVS
Message-ID: <1994Feb16.102647.17314@rahman.ox.ac.uk>
Originator: keith@rahman.ox.ac.uk
Organization: Dept of Earth Sciences, Oxford University, UK.
References: <1994Feb15.180110.109163@ucl.ac.uk>
Date: Wed, 16 Feb 1994 10:26:47 GMT
Lines: 20

In article <1994Feb15.180110.109163@ucl.ac.uk>,
Timothy Slidel <slidel@bsm.biochemistry.ucl.ac.uk> wrote: 

>Is any effort being put into converting AVS to ANSI C.  As an ANSI C
>programmer I find it a little irritating that all the AVS examples
>are still in k&r C, and that there's no type checking when sending
>arrays and stuff to AVS routines. 

A quick check through the header files in $avs_path/include reveals
that the declarations *do* have function prototypes for ANSI C and
C++, with backward compatibility for K&R C.

Are you running an old version of AVS?  We have 5.01.

Keith Refson
-- 
------------------------------------------------------------------------------
|  Email   : keith@earth.ox.ac.uk       |   Dr Keith   Refson                |
|  UUCP    : ...!mcsun!uknet!ed!K.Refson|   Department of Earth Sciences     |
| PHONE(FAX): +44 865 272026 (272072)   |   Parks Road, Oxford OX1 3PR, UK   |


From avs@doppler.NCSC.org (International AVS Center)
Subject: Re: Projected Vectors
Message-ID: <1994Feb22.152031.25436@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: avs@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Tue, 22 Feb 1994 15:20:31 GMT

Tim Rozmajzl (tim@osc.EDU) recently posted:

> I'm searching for a module that will, given a 2-D orthogonal
> slice of a 3-D irregular 3-Vector field, draw vector arrows
> projected onto the irregular 2-D slice.  Anyone tried it?

I think this module might do the trick:

Name        : show_flow       Version      : 1.100     Mod Number : 1195
Author      : Phil McDonald, NOAA/ERL/Forecast Systems Laboratory
Submitted   : 05/06/92        Last Updated : 04/05/93  Language   : C
Module path : avs.ncsc.org:avs_modules/mappers/show_flow
Ported to   : Sun IBM Kubota
Description : The show flow module generates a geometry object named
              "flow symbols" that depicts the flow in a field of 1-, 2-, or
              3-vectors of any data type and in any grid configuration.
              Show flow provides the same basic function as hedgehog but
              it has been enhanced to serve the meteorolog- ical
              community where wind barbs, rather than arrows, may be the
              desired flow symbol. Flow speed values are the resultant
              values of the field n-vector nalues. The symbols may be
              rotated about their reference points and their aspect with
              respect to the viewer can be frozen. While symbols of either
              type can be arbitrarily sized, arrows can be further scaled
              by a divisor which is the maximum of the field, the average of
              the field, or entered manually. The symbols will be colored
              by their speeds or the values in a second, optional field if
              an optional colormap is supplied.

Good luck with it.

-Steve

PS:  Would you be interested in sharing your AVS work with the
scientific visualization community worldwide via a short article 
and / or slides in an upcoming issue of AVS Network News?  This 
is the IAC's quarterly magazine made up of user contributed 
articles.   Also, the AVS '94 Program Committee 
(avs94pc@avsusers.ncsc.org) is seeking submissions for the
AVS '94 Video Review and AVS '94 Poster session.

PPS: The IAC is supported in large part by the dues it
receives from memberships at $36.00 per year.  This
includes quarterly publications, $50.00 discounts at
the conference, module portings, etc.  If you are not
already a member, please consider joining - contact 
avs@ncsc.org for more information.

Thanks for considering this!
----------------------------------------------------------------
   Steve Thorpe, Application Visualization System Specialist
	       International AVS Center, MCNC
PO Box 12889   3021 Cornwallis Rd, RTP, NC 27709   avs@ncsc.org
----------------------------------------------------------------


From fulco@cicg-communication.grenet.fr (Patrick Fulconis)
Newsgroups: comp.graphics.avs
Subject: Re: HELP: Write_any_Image module
Date: 21 Feb 1994 07:47:50 GMT
Organization: C.I.C.G. , Grenoble
Lines: 18
Distribution: world
Message-ID: <2k9ov6$qcv@cicg-communication.grenet.fr>
References: <1994Feb19.012557.2195@ee.surrey.ac.uk>
NNTP-Posting-Host: localhost.grenet.fr


you can find the library libim.a at ftp.sdsc.edu in the directory
/pub/sdsc/graphics/imtools/ibm_rs6000

--


+----------------------------------------------------------+
|                                                          | 
|     FULCONIS Patrick                                     |
|                                                          |
|     CICG (grenoble university computing center (France)) | 
|     scientific team                                      | 
|                                                          | 
|     email  fulco@grenet.fr                               | 
|                                                          | 
+----------------------------------------------------------+



From andreas@c3440.edvz.uni-linz.ac.at (Andreas Hausleitner)
Subject: AVS 5.0 on Sun
Message-ID: <1994Feb22.125456.1896@alijku05.edvz.uni-linz.ac.at>
Sender: news@alijku05.edvz.uni-linz.ac.at
Nntp-Posting-Host: c3.edvz.uni-linz.ac.at
Organization: Technical University Vienna/Austria
Date: Tue, 22 Feb 1994 12:54:56 GMT
Lines: 24

We have now installed AVS 5.0 on our Sun (SunOS 4.1.1).
AVS 5.0 works (one warning because of on old SunOS
occurs from the linker: 
 ld.so: warning: /usr/5lib/libc.so.2.6 has older revision than expected 8)

The Problem is, that selfmade modules from AVS 4.0 can't be recompiled
for AVS 5.0 because the linker terminated with

	ld: Undefined symbol 
 	  ___ansi_fflush 
	*** Error code 1

Does anyone know this problem, or are there any hints to fix this.

Thanks

Andreas Hausleitner

-----------------------------
Andreas Hausleitner
University of Linz / Austria
Department for Supercomputing
Tel.: +43 732 2468 203
email:	hausleitner@edvz.uni-linz.ac.at


From mcheng@dunlop.cs.wisc.edu (Michael Cheng)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: problems plotting LARGE amounts of 2D data?
Date: 22 Feb 1994 15:25:11 GMT
Organization: University of Wisconsin, Madison -- Computer Sciences Dept.
Lines: 26
Distribution: world
Message-ID: <2kd84n$aac@daffy.cs.wisc.edu>
NNTP-Posting-Host: dunlop.cs.wisc.edu


Hi

I am trying to find ways to plot LARGE amounts of 2D data,
and I like to know what is currently the state of the art. For the
sake of this posting, let's say "large" means much more data than
fits in real memory. From my own experience, using the virtual memory
of the workstation to store large amounts of data impedes performance
due to excessive paging.  Here is what I have been able to gather so far:

1) AVS: has 25,000 point limit. Everything is loaded into virtual memory.

2) Khoros 1: loads everything into virtual memory.
	Any updates from Khoros 2.0?

3)idl/pvwave: As far as I can tell from the short demo,
	loads everything into virtual memory?

I'm posting this to the various Comp.graphics.* 
groups, hoping that I can get feedback from users of various software
packages.  I would appreciate comments on the current/future capabilities of 
the above packages. I would also like comments about other
packages, such as SGI Explorer, IBM Data Explore, apE, or any other
package. Thanks in advance.

Mike


From cdodge@edvs11.awi-bremerhaven.de (Chris Dodge,C 0325,269,-)
Subject: Re: AVS 5.0 on Sun
Message-ID: <1994Feb22.152421.25993@news.dkrz.de>
Sender: news@news.dkrz.de (News Administrator)
Reply-To: cdodge@edvs11.awi-bremerhaven.de
Organization: Sun Microsystems
References: <1994Feb22.125456.1896@alijku05.edvz.uni-linz.ac.at>
Date: Tue, 22 Feb 94 15:24:21 GMT
Lines: 32

In article 1896@alijku05.edvz.uni-linz.ac.at, andreas@c3440.edvz.uni-linz.ac.at (Andreas Hausleitner) writes:
>We have now installed AVS 5.0 on our Sun (SunOS 4.1.1).
>AVS 5.0 works (one warning because of on old SunOS
>occurs from the linker: 
> ld.so: warning: /usr/5lib/libc.so.2.6 has older revision than expected 8)
>
>The Problem is, that selfmade modules from AVS 4.0 can't be recompiled
>for AVS 5.0 because the linker terminated with
>
>	ld: Undefined symbol 
> 	  ___ansi_fflush 
>	*** Error code 1
  
This problem is mentioned in the release notes for AVS R5.01 on SunOS 4.1.x. The soloution given in the notes is:

"In order to avoid this problem a dummy __ansi_fflush (note 2 leading underscores) function should be created and linked with modules that exhibit this linking 
problem."
 
Good luck...


  Chris Dodge    Alfred-Wegener-Institut for Polar and Marine Research
                 Postfach 12 01 61
                 27515 Bremerhaven
                 GERMANY
      
         Tel:    +49 (0)471 4831 - 269  
      E-Mail:    cdodge@awi-bremerhaven.de






From rosenbe2@ccfsun.nrl.navy.mil (Robert O. Rosenberg [Rosenberg2] )
Subject: IEEE Visualization '94 Call for Participation
Message-ID: <CLMvLC.F6L@ra.nrl.navy.mil>
Sender: usenet@ra.nrl.navy.mil
Organization: Naval Research Lab, Washington, DC
Date: Tue, 22 Feb 1994 16:10:24 GMT
Lines: 401

IEEE Visualization '94 Call for Participation

Sponsored by IEEE Computer Society Technical Committee on
Computer Graphics
In Cooperation with ACM/SIGGRAPH

October 17-21 1994

* Sheraton Premiere at Tysons Corner * Washington, DC

Scientific visualization is an important research and applications
frontier shared by a variety of science, medicine, and engineering
fields.  This conference focuses on interdisciplinary methods and
supports collaboration among the developers and users of
visualization methods across all of science, engineering, medicine and
commerce.

IEEE Visualization '94 is the fifth annual IEEE Visualization
Conference Week and will be held in Tysons Corner, in the
Washington, D.C. metropolitan area west of Arlington. The
Conference Week includes tutorials, symposium and mini-workshops
Sunday through Tuesday, and three-way parallel tracks of papers,
panels, and case studies Wednesday through Friday.

Important Visualization '94 Dates:

Paper, Panel, Case Study, Tutorial, Mini-Workshop, and BOF
Proposals Due March 31, 1994 Demonstration Proposals Due  May
30, 1994

Acceptance Notification Mailed June 10, 1994

Final Paper (Papers, Panels and Case Study) Due July 10, 1994

Tutorial Notes Due  August 10, 1994

Final Video (Paper and Case Study) Due  August 10, 1994


Paper Submissions (due March 31, 1994)

Papers are solicited that present research results related to all areas of
visualization, including visualization tools and methods, and
discipline-specific applications. Original papers should be limited to
5,000 words. The submission of images and/or NTSC video to
accompany the paper is recommended. Please submit 5 copies of all
materials. The paper will be included in the conference proceedings;
the video may be included in the conference videotape.
Contact:
    R. Daniel Bergeron
    Department of Computer Science
    University of New Hampshire
    Durham, NH 03824
    Work: (603) 862-2677
    Fax:  (603) 862-3493
    Email: rdb@cs.unh.edu


Panel Proposals (due March 31, 1994)

Panels should address the most important issues in visualization today,
with emphasis on research, applications, systems and results. Panelists
should be experts in their field who discuss the challenges of
visualization.   Summaries of panelists' position statements will
appear in the proceedings. Panel sessions are 90 - 120 minutes in
length, with 3 to 5 speakers in addition to the Panel Chair. At the
Panel Session, panelists should each make a brief position statement
addressing issues and include a moderated discussion which may
include the audience. Panel proposals should describe the topic to be
addressed and identify the prospective panelists. In the panel proposal
each panelist should include a position statement on the topic and a
short biography, the total of which should be limited to 500 words.
The statement will be included in the conference proceedings.
Contact:

    Lloyd Treinish
    IBM Thomas J. Watson Research Center
    P.O. Box 704, SK-Y57
    Yorktown Heights NY 10598
    FedX:
    IBM Research, 8 Skyline Drive, Hawthorne, NY 10532
    Work:(914) 784-5038
    Fax: (914) 784-5077
    Email: lloydt@watson.ibm.com


Interdisciplinary Case Studies (due March 31, 1994):

Case studies is a forum for scientists from various disciplines to
discuss applications, experience, and challenges of visualization, and
to present potential topics of future research. These sessions provide
an interdisciplinary meeting point between scientists, engineers and
physicians from different areas such as astrophysics, atmospheric
sciences, computational fluid dynamics, engineering, geology,
medicine, anthropology, chemistry, etc.  Please submit a short paper
(1000 words) describing the case study. The submission of images
and/or NTSC video to accompany the paper is recommended. A short
paper (maximum 4 pages including images) will be included in the
conference proceedings; the video may be included in the conference
videotape.
Contact:

    Deborah Silver
    Dept of Electrical and Computer Engineering
    CAIP Center
    Rutgers University
    P.O. Box 1390
    Core Building, Frelinghuysen Road
    Piscataway NJ 08855-1390
    Work: (908) 932-5546
    Fax:  (908) 932-4775
    Email: silver@caip.rutgers.edu


Mini-Workshop and Birds-of-a-Feather (BOF) Proposals
(due March 31, 1994)

Evening mini-workshops and birds-of-a-feather (BOF) gatherings on
specific visualization methods or application areas will be offered on
Sunday or as evening sessions.  They  should deal with state-of-the-art
topics and involve experts in the field. Discipline-focused proposals
devoted to a particular discipline's methods and needs are encouraged
(e.g. visualization related to molecular biology). If appropriate, the
session may be co-sponsored by another professional organization.
These sessions are forums for participants to exchange views. Mini-
workshop organizers select approximately 20 participants based on
responses to a CFP published in the Visualization '94 Advance
Announcement. BOFS will be described in the Advance Program and
Final Program and attendance is ad hoc. Proposals should describe
the theme and goals of the Mini-Workshop or BOF, the activities
planned, and a brief description of the organizer's background.
Contact:

    Chuck Hansen
    Los Alamos National Laboratory
    MS-B287
    P.O. Box 1663
    Los Alamos, NM 87545
    Work: (505) 665-3663
    Fax:  (505) 665-4939
    Email: hansen@lanl.gov


Tutorial Proposals (due March 31, 1994)

Half-day and full-day course proposals are invited for visualization
systems, methods, and application areas. Tutorials will be offered
Sunday, Monday and Tuesday, preceding the Wednesday through
Friday Visualization '94 Conference.  Proposals should clearly
identify the visualization proficiencies expected of participants at a
beginning, intermediate, or advanced level. Tutorial proposals should
include an abstract of the tutorial topic, a description of the tutorials'
organization to include time allocation for major course topics, the
duration (1/2 day or 1 day), the level of the tutorial, and the
background and address information of each of the instructors.
Tutorial proposals will be evaluated on the basis of the tutorial's
relevance, importance and suitability for presentation in a tutorial
format; the past experience and qualifications of the instructors, and
the overall balance in the tutorial program.  The Tutorial Organizer
will be responsible for development of the Tutorial Notes for the
Tutorial, the notes will be due in August.
Contact:

    Hikmet Senay
    Department of Electrical Engineering and Computer Science
    The George Washington University
    Washington, D.C. 20052
    Phone: (202) 994-5910
    Fax:   (202) 994-0227
    Email: senay@seas.gwu.edu


Demonstration Proposals (due May 30, 1994)

Visualization '94 is a unique opportunity to present your products to
visualization experts from a wide variety of fields. We invite
applications for demonstrations of commercial hardware, software,
integrated systems, peripherals, and literature. We encourage
commercial demonstrators to have technical representatives in
attendance. Call or write for a packet which includes fees for
commercial demonstrations and past attendee demographics.
Registration for the technical conference is included in the
commercial demonstration fee. Commercial demonstrators confirmed
before June 15, 1994 will be announced in the Visualization '94
Advance Program.

Research groups from academia and research labs are invited to
demonstrate their work interactively at Visualization '94.  Proposals
should summarize the work to be presented and identify the
hardware/software platform required. Proposals from non-profit
organizations will be reviewed and will be accepted based on the
anticipated level of interest in the research and the space available.
Demonstrations will be held on Wednesday and Thursday during the
conference.
Contact:

    Sally Wood
    Department of Electrical Engineering
    Santa Clara University
    Santa Clara, CA
    Work: (408) 554-4058
    Fax:  (408) 554-5474
    Email: swood@scuacc.scu.edu


Call for Corporate Partners

We are interested in organizations to partner with the Visualization
'94 Conference Committee in offering this highly technical
visualization week.  Visualization '94 is a non-profit organization and
the support of Corporate Partners allows the conference to keep
registration costs to a minimum, while providing state-of-the art
information about visualization for attendees from a novice
implementor to an advanced researcher level.  Corporate Partners are
acknowledged in the Advance and Final Programs, and in the
conference proceedings. For more information, please contact Nahum
Gershon, (703) 883-7518 or email gershon@mitre.org

Visualization '94 Conference Co-Chairs:

Nahum Gershon, The MITRE Corporation, (703) 883-7518, email
gershon@mitre.org

Carol Hunter, Lawrence Livermore National Laboratory, (510) 422-
1657, email chunter@llnl.gov

Visualization '94 Program Co-Chairs:

Larry Rosenblum, Office of Naval Research European Office, email
lrosenblum@onreur.navy.mil

Bill Ribarsky, Georgia Institute of Technology, (404) 894-6148,
email bill.ribarsky@oit.gatech.edu

IEEE Visualization 94 Proposal/Speaker Form

Please fill out a speaker form for each proposal you submit. For
proposals involving multiple presenters (e.g. panel, tutorial) please
fill out a speaker form for each presenter)
Speaker's
Name:___________________________________________________

Affiliation:
_________________________________________________________

Mailing
Address:_________________________________________________
_________________________________________________________

_________________________________________________________

City:______________State:_____ Country:_________Zip:_________


Phone: __________________________
Fax:_________________________________

Email:___________________________________________________

Circle Type of Session:  paper    panel    case study   mini-workshop

tutorial   BOF   demonstration    Volume Visualization Symposium
Title of
Presentation:______________________________________________

_________________________________________________________

Please list names, addresses, phone and fax numbers, and e-mail
addresses of the additional authors/presenters:
_________________________________________________________



IEEE Visualization '94 Conference Committee

Honorary Visualization '94 Conference Chair
    Andries van Dam, Brown University

Conference Co-Chairs
    Nahum Gershon, The MITRE Corporation
    Carol Hunter, Lawrence Livermore National Laboratory

Program Co-Chairs
    Bill Ribarsky, Georgia Institute of Technology
    Larry Rosenblum, Office of Naval Research, European Office

Papers Co-Chairs
    Dan Bergeron, University of New Hampshire
    Arie Kaufman, State University of New York at Stony Brook

Panels Co-Chairs
    Lloyd Treinish, IBM T.J. Watson Research Center
    Kevin Hussey, Jet Propulsion Laboratory

Case Studies Co-Chairs
Deborah Silver, Rutgers University
Chuck Goodrich, University of Maryland

Tutorials Co-Chairs
    Hikmet Senay
    Greg Nielson, Arizona State University

Mini-Workshops and BOFS Co-Chairs
    Chuck Hansen, Los Alamos National Laboratory
    Polly Baker, National Center for Supercomputing applications

Video Proceedings Co-Chairs
    Robert McDermott, University of Utah
    James Rose, University of Utah

CDROM Co-Chairs
    Steve Follin, University of Georgia
    Tony Scarlatos, State University of New York at Stony Brook

Demonstrations Co-Chairs
    Sally Wood, Santa Clara University
    Ed Council, Timberfield Systems
    Susan Stearman, Maven Consultants

Publicity Co-Chairs
    Theresa-Marie Rhyne, Martin Marrietta/EPA
Georges Grinstein, University of Massachusetts at Lowell
Carol Hunter, Lawrence Livermore National Laboratory

Registration Co-Chairs
    Ross Gaunt, Lawrence Livermore National Laboratory

Student Volunteers  Co-Chairs
Stephen Watson, Jet Propulsion Laboratory
JP Lee, University of Massachusetts at Lowell

Audio Visual Co-Chairs
Len Wanger, San Diego Supercomputer Center

Local Co-Chairs
Upul Obeysekare, Naval Research Laboratory
Robert Rosenberg, Naval Research Laboratory
Janet Jensen, ERDEC
    Larry Schuette, Naval Research Laboratory
    Ray Twiddy, Hughes STX
    John Hagedorn, SSAI
    Ken Musgrave, George Washington University
    Margaret Douglas, NIH
    Nancy Johnston, Lawrence Berkeley Laboratory

Finance Co-Chairs
    Michael Danchak, Hartford Graduate Center
    Shirley Stephan, Lawrence Livermore National Laboratory

International Liaison Co-Chairs
    Phil Robertson, CSIRO Information Division, Australia
    Frits Post, University of Delft, The Netherlands
    Stas Klimenko, IHEP, Russia

ACM/IEEE Volume Vizualization Symposium
    Roni Yagel, Ohio State University
    David Ebert, University of Maryland


IEEE Visualization '94 Program Committee


Mike Bailey, San Diego Supercomputer Center
H. Harlyn Baker, SRI International
Stephen Benton, MIT
Steve Bryson, CSC at NASA Ames
Susan Chipman, Office of Naval Research
Donna Cox, NCSA
Jeff Dozier, Univ. of Calif., Santa Barbara
Rae A. Earnshaw, Univ. of Leeds, UK
Steven Eick, AT&T Bell Labs
Jose Encarnacao, Technische Hochschule, Darmstadt, Germany
Steve Feiner, Columbia Univ.
Jim Foley, Georgia Institute of Technology
Henry Fuchs, Univ. of NC, Chapel Hill
Issei Fujishiro, Ochanomizu Univ., Tokyo
Richard S. Gallagher, Swanson Data Analysis
Michel Grave, ONERA, France
Hans Hagen, Universitaet Kaiserslautern, Germany
Lambertos Hesselink, Stanford University
William Hibbard, Univ. of WI - Madison
Karl Heinz H?hne,  Univ. of Hamburg
F.R.A. Hopgood, Rutherfrod Appleton, UK.
Rob Jacob, Naval Research Laboratory
Larry Gelberg, Application Visualization Systems
Fred Kitson, Hewlett Packard Labs
Stanislav Klimenko, Inst. of Physics, Russia
Tosiyasu Kunii, Univ. of Tokyo
Marc Levoy, Stanford University
Glenn Mucklow, NASA
Art Olson, Research Institute of Scripps Clinic
Mike Rhodes, Toshiba America
Azriel Rosenfeld, Univ. of Maryland
Werner Stuetzle, Univ. of Washington
Jeffery Star, Univ. of Calif, Santa Barbara
Nadia Thalmann, University of Geneva
Gary Watkins, Evans & Sutherland
Val Watson, NASA Ames
Peter Wilson, Rensselaer Polytechnic Inst.
Allan R. Wilks, AT&T Bell Laboratories
James M. Winget, Silicon Graphics, Inc.


From zawodny@arbd0.larc.nasa.gov (Joseph M Zawodny)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Problem with structures getting padded to multiples of 4-bytes
Date: 22 Feb 1994 16:24:58 GMT
Organization: NASA Langley Research Center, Hampton, VA
Lines: 113
Distribution: world
Message-ID: <2kdbkq$atb@reznor.larc.nasa.gov>
NNTP-Posting-Host: arbd0.larc.nasa.gov

Howdy All,

	I've come a cross something a bit odd with large structures. I'm
running IDL Version 3.5.1 (OSF alpha) on a DECStation 3000 Model 400 AXP.
I've used a lot of small structures that are a non-integer number of (4-byte)
words in length with little problem, however, the following structure
should be 2618 bytes in length but comes up 2620.  When written to a file,
the record is indeed 2620 bytes long and the last two (extra) bytes appear to
be padded with zero's.  Is this a known problem and are there work arounds for
it?  Rumor has it that this does not happen on a SUN SPARC-10 running whatever
they run for an OS this week.

	Thanks,

The structure:

;
; Make some temporary arrays
tmpf4  = fltarr(4)
tmpf8  = fltarr(8)
tmpf25 = fltarr(25)
tmpf70 = fltarr(70)
impf70 = intarr(70)

; Define the structure
ozn  = {OZN_STR,  $
;       Field          Type 
        DATE:           0.,     $
        TIME:           0.,     $
        LAT:            0.,     $
        LON:            0.,     $
        TYPE_INST:      0.,     $
        TYPE_LOCAL:     0.,     $
        BETA:           0.,     $
        TIMEOFYR:       0.,     $
        NMCTEMP:        tmpf25, $
        NMCTERR:        tmpf25, $
        NMCALT:         tmpf25, $
        NMCDEN:         tmpf25, $
        NMCDERR:        tmpf25, $
        METCOR:         tmpf4,  $
        MET_POINTER:    0.,     $
        MET_CODE:       0.,     $
        MET_REV_DATE:   0.,     $
        PROC_LEVEL:     0.,     $
        PROC_DATE:      0.,     $
        PROC_TIME:      0.,     $
        MEANSBTALT:     0.,     $
        MIRROR_CAL:     0.,     $
        QUALF600:       0.,     $
        FILLVAL:        0.,     $
        SBTNLAT:        tmpf8,  $
        SBTNLON:        tmpf8,  $
        TIMESPAN:       0.,     $
        GEOALT:         tmpf70, $
        PRESS:          tmpf70, $
        TEMP:           tmpf70, $
        O3ND:           tmpf70, $
        O3MR:           tmpf70, $
        O3NDE:          impf70, $
        O3MRE:          impf70, $
        ND_QUAL_FLAG:   impf70, $
        MR_QUAL_FLAG:   impf70, $
        INT_FILL:       0    }   


Doing a help,/str,ozn yields:

** Structure OZN_STR, 37 tags, length=2620:
   DATE            FLOAT           0.00000
   TIME            FLOAT           0.00000
   LAT             FLOAT           0.00000
   LON             FLOAT           0.00000
   TYPE_INST       FLOAT           0.00000
   TYPE_LOCAL      FLOAT           0.00000
   BETA            FLOAT           0.00000
   TIMEOFYR        FLOAT           0.00000
   NMCTEMP         FLOAT     Array(25)
   NMCTERR         FLOAT     Array(25)
   NMCALT          FLOAT     Array(25)
   NMCDEN          FLOAT     Array(25)
   NMCDERR         FLOAT     Array(25)
   METCOR          FLOAT     Array(4)
   MET_POINTER     FLOAT           0.00000
   MET_CODE        FLOAT           0.00000
   MET_REV_DATE    FLOAT           0.00000
   PROC_LEVEL      FLOAT           0.00000
   PROC_DATE       FLOAT           0.00000
   PROC_TIME       FLOAT           0.00000
   MEANSBTALT      FLOAT           0.00000
   MIRROR_CAL      FLOAT           0.00000
   QUALF600        FLOAT           0.00000
   FILLVAL         FLOAT           0.00000
   SBTNLAT         FLOAT     Array(8)
   SBTNLON         FLOAT     Array(8)
   TIMESPAN        FLOAT           0.00000
   GEOALT          FLOAT     Array(70)
   PRESS           FLOAT     Array(70)
   TEMP            FLOAT     Array(70)
   O3ND            FLOAT     Array(70)
   O3MR            FLOAT     Array(70)
   O3NDE           INT       Array(70)
   O3MRE           INT       Array(70)
   ND_QUAL_FLAG    INT       Array(70)
   MR_QUAL_FLAG    INT       Array(70)
   INT_FILL        INT              0



-- 
 Joseph M. Zawodny   (KO4LW)                    NASA Langley Research Center
 Internet: zawodny@arbd0.larc.nasa.gov          MS-475, Hampton VA, 23681-0001
 Packet:   ko4lw@n4hog.va.usa


From avs@NCSC.org
Subject: AVS Net News Call for Images
Message-ID: <1994Feb22.175051.29329@mcnc.org>
Sender: daemon@mcnc.org (David Daemon)
Nntp-Posting-Host: doppler.ncsc.org
Reply-To: avs@ncsc.org 
Organization: North Carolina Supercomputing Center
Date: Tue, 22 Feb 1994 17:50:51 GMT

Dear AVS Users,

The IAC has been soliciting submissions for the Spring 1994 issue
of AVS Network News, which is to hit the streets on May 2, 1994,
in conjunction with the AVS '94 Conference in Boston.

This will be a special "Image Gallery" edition of AVS Network News, 
that will contain only one or two short articles, by the IAC.  The
rest of the magazine will be devoted to FULL COLOR images submitted
by the AVS user community.

All AVS users are invited to submit 1-5 images for this issue.  
Depending on the number of submissions, the IAC may not be able 
to print all or any of an individual's submissions.

Spring '94 Net News submissions are due Monday 3/7/94 at 12 Noon EST.
                                        ----------------------------

   How to submit your images
   -------------------------
   1)  Create your 1-5 image(s) in AVS .x format.

       To help us keep organized on this end, please name
       your images using the format "FirstName_LastName.Number.x".
       So for example if I was submitting 3 images, I would name
       them Steve_Thorpe.1.x, Steve_Thorpe.2.x, and Steve_Thorpe.3.x.

   2)  Create a short ASCII file with a title, description, and
       credits for each image submitted.  

       An image title should be something short, ie "Intergalactic 
       Stellar Gas Cloud").

       Image descriptions should be no more than 40 words for each
       image.

       The credits for each image should list who
       created it and affiliation, also if applicable 
       the source of the data.

       Name this image description file using a similar style as 
       with the image names.  For example, the ASCII description
       file that I would submit with my three images would be called 
       Steve_Thorpe.txt.
	
   3)  Submit the files to the IAC using anonymous ftp.  For example,
       to submit my images I would do the following:

       ftp avs.ncsc.org             (otherwise known as 128.109.178.23)
       cd SUBMIT
       mkdir Steve_Thorpe_Images    (substitute in your name)
       cd Steve_Thorpe_Images       (substitute in your name)
       ascii
       put Steve_Thorpe.txt
       binary
       put Steve_Thorpe.1.x         (substitute in your name)
       put Steve_Thorpe.2.x         (substitute in your name)
       put Steve_Thorpe.3.x         (substitute in your name)
       quit

   4)  Print out a copy of the "AVS Network News Permission to Use 
       Images Form" (included at the end of this message), fill it out
       and FAX the complete form to Steve Thorpe at 919-248-1101.

Please feel free to give me a call at (919) 248-1161
if you have any questions or thoughts about AVS Network News.

Thank you very much for your interest in contributing.  Your work
can make a very valuable contribution to the Scientific Visualization 
community worldwide. 

And also, please remember that Spring '94 Net News submissions 
are due Monday 3/7/94 at 12 Noon EST.  
        ----------------------------

-Steve
----------------------------------------------------------------
   Steve Thorpe, Application Visualization System Specialist
International AVS Center, North Carolina Supercomputing Center
PO Box 12889   3021 Cornwallis Rd, RTP, NC 27709   avs@ncsc.org
----------------------------------------------------------------

----------------------------------------------
AVS Network News Permission to Use Images Form
----------------------------------------------

Complete this form if you are submitting any images 
for AVS Network News and FAX it to Steve Thorpe at
919-248-1101.

A copy of this form MUST accompany each submission.  Submit
this along with the image(s) and image description file.
 
Please keep photocopy for your records.  

   Name of submitter:
        Organization:
      Address Line 1:
           Address 2:
           Telephone:
                 Fax:
               Email:

File Name of image 1:
    Title of image 1:

File Name of image 2:
    Title of image 2:

File Name of image 3:
    Title of image 3:

File Name of image 4:
    Title of image 4:

File Name of image 5:
    Title of image 5:

                AUTHORIZATION

I grant permission for the above-referenced material 
to be published in the International AVS Center's quarterly
magazine AVS Network News.  In the event that any materials 
used contain the work of other individuals or organizations 
including any copyrighted materials, data, or excerpts thereof, 
I understand that it is my responsibility to secure any necessary 
permissions.  By submitting this form I hereby indicate that all
such permissions have been obtained:

o I have the necessary rights and/or permissions described above.

o I grant non-exclusive worldwide distribution rights to
  the International AVS Center to publish and distribute my images
  with appropriate credits being attached


Signature___________________________

Date________________________________


STATEMENT OF PURPOSE:  The AVS Network News Permission to Use 
Images Form explains the uses the International AVS Center plans
to make of your submitted images and credits, and requires you 
to acknowledge that you have permission to use this material.  
This may involve seeking clearance from your employer or from 
others who have loaned you material, such as videotapes and slides.  
This form helps prevent situations where AVS Network News 
includes material without permission that might lead to complaints 
or even legal action.  This form must be returned to the International
AVS Center as a condition for acceptance for publication in AVS Network
News.  



From tim@osc.EDU (Tim Rozmajzl)
Newsgroups: comp.graphics.avs
Subject: UCD module
Date: 22 Feb 1994 17:42:12 GMT
Organization: Ohio Supercomputer Center
Lines: 27
Distribution: world
Message-ID: <2kdg5k$jou@mane.cgrg.ohio-state.edu>
NNTP-Posting-Host: killian.osc.edu


Hello,

I've recently created a module that reads a data file and outputs
a UCD structure.  It's written in FORTRAN.  It seems to read the
file without problems.  ucd_to_geom receives the output from my
module and generates the expected geometry.  However, ucd_contour
fails when it receives the UCD structure from my module.  (fatal
FLOWEXEC error).  

The strange thing is that if I pipe the output from my module 
directly to write_ucd, then read that file back in using read_ucd,
ucd_contour works fine.  Has anyone encountered this type of 
behavior before?  Any solutions?

Thanks in advance,

Tim

-- 
o------------------------The Ohio Supercomputer Center ---------------o
o  Tim Rozmajzl                   |  Phone: (614) 292-3105            o
o  User Services		  |  FAX:   (614) 292-7168            o
o  The Ohio Supercomputer Center  |  Email: tim@osc.edu               o
o  1224 Kinnear Rd.               |                                   o
o  Columbus, Ohio 43212           |                                   o
o---------------------------------------------------------------------o


From thompson@serts.gsfc.nasa.gov (William Thompson)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Re: problems plotting LARGE amounts of 2D data?
Date: 22 Feb 94 17:22:48 GMT
Organization: NASA Goddard Space Flight Center -- InterNetNews site
Lines: 44
Distribution: world
Message-ID: <thompson.761937768@serts.gsfc.nasa.gov>
References: <2kd84n$aac@daffy.cs.wisc.edu>
NNTP-Posting-Host: serts.gsfc.nasa.gov
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

mcheng@dunlop.cs.wisc.edu (Michael Cheng) writes:


>Hi

>I am trying to find ways to plot LARGE amounts of 2D data,
>and I like to know what is currently the state of the art. For the
>sake of this posting, let's say "large" means much more data than
>fits in real memory. From my own experience, using the virtual memory
>of the workstation to store large amounts of data impedes performance
>due to excessive paging.  Here is what I have been able to gather so far:

>1) AVS: has 25,000 point limit. Everything is loaded into virtual memory.

>2) Khoros 1: loads everything into virtual memory.
>	Any updates from Khoros 2.0?

>3)idl/pvwave: As far as I can tell from the short demo,
>	loads everything into virtual memory?

>I'm posting this to the various Comp.graphics.* 
>groups, hoping that I can get feedback from users of various software
>packages.  I would appreciate comments on the current/future capabilities of 
>the above packages. I would also like comments about other
>packages, such as SGI Explorer, IBM Data Explore, apE, or any other
>package. Thanks in advance.

>Mike

I think the way you're looking at this is all wrong.  It's not AVS/Khoros/IDL
which decide whether or not data are in real or virtual memory--it's the
operating system.  Operating systems such as Unix or VMS shield the program
from knowing how much of the memory they're using is real or virtual.
Generally speaking, you'll get as much of the real memory as possible and only
the overflow will be stored in virtual memory.  Any programs which are designed
to run on a virtual memory operating system should behave the same way.

If a program such as IDL (or any of the others you mention) is putting things
into virtual memory, it can only be because you don't have enough real memory
available to you.  The only thing you can do is buy more memory, or if your
operating system supports memory usage quotas (such as VMS) then you need to
increase your quotas.

Bill Thompson


From thompson@serts.gsfc.nasa.gov (William Thompson)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Re: Problem with structures getting padded to multiples of 4-bytes
Date: 22 Feb 94 17:38:25 GMT
Organization: NASA Goddard Space Flight Center -- InterNetNews site
Lines: 63
Distribution: world
Message-ID: <thompson.761938705@serts.gsfc.nasa.gov>
References: <2kdbkq$atb@reznor.larc.nasa.gov>
NNTP-Posting-Host: serts.gsfc.nasa.gov
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

zawodny@arbd0.larc.nasa.gov (Joseph M Zawodny) writes:

>Howdy All,

>	I've come a cross something a bit odd with large structures. I'm
>running IDL Version 3.5.1 (OSF alpha) on a DECStation 3000 Model 400 AXP.
>I've used a lot of small structures that are a non-integer number of (4-byte)
>words in length with little problem, however, the following structure
>should be 2618 bytes in length but comes up 2620.  When written to a file,
>the record is indeed 2620 bytes long and the last two (extra) bytes appear to
>be padded with zero's.  Is this a known problem and are there work arounds for
>it?  Rumor has it that this does not happen on a SUN SPARC-10 running whatever
>they run for an OS this week.

>	Thanks,

>The structure:

	(rest deleted)

I've run into this sort of thing myself.  A much simpler structure that
exhibits the same behavior is the following:

	TEST = {FLOAT: 0.0, BYTE: 0B}

This should, in principal, be five bytes long, but comes out as being eight
bytes instead.  The behavior is consistent on a Sparcstation 1, on a DEC
5000/240 and on an AXP 3000/400 running OpenVMS.  Apparently, structures want
to be an even number of "words" (four bytes) long.

Another factor to consider is the following.  Consider the structure:

	TEST2 = {FLOAT: 0.0, BYTE1: 0B, INTEGER: 0, BYTE2: 0B}

This should be eight bytes long, but instead it comes out as 12.  However, by
slightly modifying the structure to

	TEST3 = {FLOAT: 0.0, BYTE1: 0B, BYTE2: 0B, INTEGER: 0}

then it comes out to 8 as it should.  Apparently, integers want to be start on
"half-word" (two bytes) boundaries.

I use the terms "word" and "half-word" in quotes because, as far as I
understand it, the true wordsize on AXP platforms should be eight bytes.
However, IDL on AXPs doesn't seem to behave any different than on any other
platform.

These considerations are important when doing reads or writes with structure
variables.  The only workarounds that I can suggest is either

	1.  Design your structures carefully to conform to the above
	    restrictions, or

	2.  Read the data in as byte arrays, and then use the IDL data
	    conversion functions--FIX, FLOAT, DOUBLE, COMPLEX, STRING--to
	    convert this into a structure, or visa-versa for writes.

There is some information on this in the IDL manual, under "Unformatted
Input/Output with Structures" in the chapter on structures.

Good luck,

Bill Thompson


From zaxin@bigwpi.WPI.EDU (Benjamin C. Lee)
Newsgroups: comp.graphics.avs
Subject: Iges CAD files
Date: 22 Feb 1994 21:12:13 GMT
Organization: Worcester Polytechnic Institute
Lines: 20
Message-ID: <2kdsfd$k72@bigboote.WPI.EDU>
NNTP-Posting-Host: bigwpi.wpi.edu
Keywords: avs iges cad


	I was wondering, are there any utilities around that would aid
in a translation from the Iges format for CAD type things to the AVS
geometry format?

	It would be nice if there was a utility or module that could
do it for me, but I am not totally closed to the idea of writing one
myself.  In which case I'll need to know the Iges format.

	So, I would love to hear from anyone who has heard of such a
utility/module, are has any knowledge of the Iges format to offer.

					thanks


-- 
--
Disclaimer: I don't even know her!  | Benjamin "It's the hole I've dug" Lee
Nothing in Moderation; Do what you do... Do it well  |  zaxin@wpi.wpi.edu
GCS -d+(---) p c++++ l u++ e+/* m--- s+++/+ n- h* f+(*) g++ w+++ t--- r+ y*


From mcheng@dunlop.cs.wisc.edu (Michael Cheng)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Re: problems plotting LARGE amounts of 2D data?
Date: 22 Feb 1994 22:51:12 GMT
Organization: University of Wisconsin, Madison -- Computer Sciences Dept.
Lines: 27
Distribution: world
Message-ID: <2ke290$fac@daffy.cs.wisc.edu>
References: <2kd84n$aac@daffy.cs.wisc.edu> <thompson.761937768@serts.gsfc.nasa.gov>
NNTP-Posting-Host: dunlop.cs.wisc.edu

In article <thompson.761937768@serts.gsfc.nasa.gov> thompson@serts.gsfc.nasa.gov (William Thompson) writes:

>If a program such as IDL (or any of the others you mention) is putting things
>into virtual memory, it can only be because you don't have enough real memory
>available to you.  The only thing you can do is buy more memory, or if your
>operating system supports memory usage quotas (such as VMS) then you need to
>increase your quotas.

I argue that buying more memory is not always the best solution for the
following reasons:
1) Some of us are poor. (arguably a weak reason)
2) Some data are always larger than the largest amount of memory 
reasonable amounts of money can buy.
3) Even a few medium sized data can overload real memory quickly.
For example, working with five 25-meg data sets already requires 125 megs
of memory.

There has always been a mistmatch between virtual memory policy and
the need to handle large amounts of data. This mismatch has been demonstrated
time and again in database systems. I feel that this issue will come
up again with respect to scientific data sets.

So we go back to my original question: is there any software package
for plotting large amounts of 2D data that does better than loading
everything into virtual memory?

Mike


From peter@hpl.hp.com (Peter Webb)
Subject: Re: problems plotting LARGE amounts of 2D data?
Sender: news@hpl.hp.com (HPLabs Usenet Login)
Message-ID: <CLnoL2.nD@hpl.hp.com>
Date: Wed, 23 Feb 1994 02:36:38 GMT
References: <2kd84n$aac@daffy.cs.wisc.edu>
Nntp-Posting-Host: oz.hpl.hp.com
Organization: Hewlett-Packard Laboratories, Palo Alto, CA
X-Newsreader: TIN [version 1.2 PL2]
Followup-To: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Lines: 41

Michael Cheng (mcheng@dunlop.cs.wisc.edu) wrote:

: Hi

: I am trying to find ways to plot LARGE amounts of 2D data,
: and I like to know what is currently the state of the art. For the
: sake of this posting, let's say "large" means much more data than
: fits in real memory. From my own experience, using the virtual memory
: of the workstation to store large amounts of data impedes performance
: due to excessive paging.  Here is what I have been able to gather so far:

: 3)idl/pvwave: As far as I can tell from the short demo,
: 	loads everything into virtual memory?

IDL and PV-Wave have a 'associated' variable type, which lets you 
map an array onto a disk file, and then step through the disk file
one slice at a time.  For example, 

	openr, lun, 'myfile.dat', /get_lun
	data = assoc(lun,bytarr(1024,1024))

associated the variable name data with the contents of file myfile.dat.
Nothing has been read yet.  Then the statement

	slice = data(10)

will load the 11th bytarr(1024,1024) contained in the file into working 
memory.  A statement like

	data(10) = byte(fft(data(10)))

will read, process, and replace the 11th array in the disk file.

So, if you can access data in a nice order, so that you don't have to 
go out to disk all the time, you can work with enormous data sets.

Something like this should also work for plotting a 2D data set.

Hope this helps.

 - Peter


From mayer@piis05.joanneum.ac.at (Harald Mayer)
Newsgroups: comp.graphics.avs
Subject: Re: ANSI AVS
Date: 23 Feb 1994 07:51:27 GMT
Organization: Joanneum Research, Graz, Austria
Lines: 34
Message-ID: <2kf1tvINNhrl@fstgds15.tu-graz.ac.at>
References: <1994Feb15.180110.109163@ucl.ac.uk> <1994Feb16.102647.17314@rahman.ox.ac.uk>
NNTP-Posting-Host: piis05.joanneum.ac.at
X-Newsreader: TIN [version 1.2 PL2]

Keith Refson (keith@earth.ox.ac.uk) wrote:
: A quick check through the header files in $avs_path/include reveals
: that the declarations *do* have function prototypes for ANSI C and
: C++, with backward compatibility for K&R C.

: Keith Refson

Yes, but there is one function prototype in <avs/field.h> for AVSbuild_field
(very old unsupported function) which looks like

	AVS_EXTERN( AVSfield *AVSbuild_field, (...));

and expands to
	
	extern AVSfield *AVSbuild_field(...);

and this syntax is not compilable with the standard compilers shipped
with DEC computers (MIPS cc, DEC C c89). In my opinion this construction
is not ANSI C (there has to be at least one fixed argument in this list).

So if AVS is going to suppurt ANSI C, which is *very* important, they should
be careful to use correct prototypes and check it with all compilers.
I think the normal AVS user/programmer should not have to change AVS
include files.

Harald
--
------------------------------------------------------------------------------
 Harald Mayer                      | X.400:    <S=mayer;P=joanneum;A=ada;C=at>
 Joanneum Research                 | Internet: mayer@pbox.joanneum.ac.at
 Institute for Information Systems | VoiceNet: +43 (316) 8020 - 136
 Steyrergasse 17       	           | FaxNet:   +43 (316) 8020 - 191
 A-8010 Graz, Austria              |
------------------------------------------------------------------------------


From ICH561@DJUKFA11.BITNET (Astrid Kuhr)
Newsgroups: comp.graphics.avs
Subject: 3D-cloud visualization searched
Message-ID: <94054.084150ICH561@DJUKFA11.BITNET>
Date: 23 Feb 94 07:41:50 GMT
Organization: Forschungszentrum Juelich
Lines: 36


Hello!

Is there anywhere a solution for the visualization (with AVS) for the
following problem??
We have a lot of measurepoints, in 3D, each point with a x,y,z coordinate.
At these points we are measurering a concentration. So we have a set of
(x,y,z,value) points.
The problem is, that the distribution of x,y,z is not on a grid. They are
somewhere 'in the air' (in the ground)
How can I visualize this cloud???
Does anybody has a module, which can do this??
Any help is welcome and very urgent needed!!

If somebody is interested in the background: we have boreholes, their
location give the x,y coordinate. In the borehole their is a defined place,
where we get our probe, this is the z-coordinate. From the probe
(waterprobe) we determine the concentration of a special substance.
This is our 'value'. We are measuring at different times (but thats not
the problem) and want to visualize the flow of the substance-cloud through
the field.
What in makes complicated for me and not solveable by myself: the location
of the boreholes is somewhere at the field, not following a determined
sheme.

Very thanx in advance for every help!
And please excuse my not so good english.


Regards, Astrid Kuhr

--
Astrid Kuhr                                 Email: a.kuhr@kfa-juelich.de
ICG-4, Institute of chemistry and dynamics  Fax:   (+49)-(0)2461-61-2484
of the geosphere
Research Centre Juelich (KFA), Germany (W)


From mccau002@acpub.duke.edu (Gary McCauley)
Newsgroups: comp.graphics.avs
Subject: Saving xy location for hand drawn contours?
Date: 23 Feb 1994 14:38:08 GMT
Organization: Duke University; Durham, NC; USA
Lines: 17
Message-ID: <2kfpog$qd8@news.duke.edu>
NNTP-Posting-Host: north8.acpub.duke.edu


  I would like to be able to manually draw a contour on an image
  and save the point locations for the points which comprise the
  contour in order to describe them as polygons for 3d reconstruction.

  Is there a way to do this? If so, how? I have figured out how to
  use sketch_roi but I do not know how (if possible) to save the
  drawn ROI boundary point locations.

  Any input would be appreciated.


  gary mccauley
 
  please respond here or to
  mccauley@watt.radonc.duke.edu



From slidel@bsm.bioc.ucl.ac.uk (Timothy Slidel)
Subject: Re: ANSI AVS
Message-ID: <1994Feb23.160751.109117@ucl.ac.uk>
Date: Wed, 23 Feb 1994 16:07:51 GMT
References: <1994Feb15.180110.109163@ucl.ac.uk> <1994Feb16.102647.17314@rahman.ox.ac.uk>
Organization: University College London
Lines: 47

In article <1994Feb16.102647.17314@rahman.ox.ac.uk>, keith@earth.ox.ac.uk (Keith Refson) writes:
|> A quick check through the header files in $avs_path/include reveals
|> that the declarations *do* have function prototypes for ANSI C and
|> C++, with backward compatibility for K&R C.
|> 
|> Are you running an old version of AVS?  We have 5.01.

We still have AVS 5.0 (soon 5.01) - the AVS headers do have the ANSI and C++
prototypes though. However, a paragraph in the release notes for our SG platform,
section "Release Notes 3-15" reads:

"As an examination of the Makeinclude file would show, when compiling modules
written in C you must use the cc compiler with the -cckr option. This option
uses a C language standard as defined in the Kernighan and Ritchie C book, rather
than ANSI C. AVS requires the somewhat greater leniency afforded by -cckr."

Is this a fallacy ? - if so it shouldn't be in the release notes - which are
supposed to be more up-to-date than the ref. manuals.

The vast majority of the source code examples that come with AVS 5 are written
in K&R C - only a few ANSI/C++ examples are given and these have .cpp extensions !
ANSI was adopted ~6 years ago - surely there should be more ANSI examples than
K&R ones. I appreciate that many people in the scientific community still use
K&R C - but when using something like AVS the strong(er) type checking, prototyping
etc. in ANSI C is essential. This may seem pedantic but the examples are the
first thing a user goes to once they've been confused by the manuals. I hope
the next release of AVS will clarify the situation - unfortunatelly, by then,
we'll all probably be using C++ !!!

-Tim-

-- 
**************************************************
Tim Slidel

Biomolecular Structure and Modelling Unit (BSM),
Dept. Biochemistry and Molecular Biology,
University College London,
Gower Street,
London WC1E 6BT.

Tel: (071) 387 7050 x3896
Fax: (071) 380 7193
Email: slidel@uk.ac.ucl.bioc.bsm
**************************************************




From rnh@uk.co.gec-mrc (Richard Herring)
Newsgroups: comp.graphics.avs
Subject: "Synergy at work" Poster
Message-ID: <4124@mule>
Date: 22 Feb 94 15:59:23 GMT
Organization: GEC-Marconi Research Centre, Great Baddow, UK
Lines: 9
X-Newsreader: TIN [version 1.2 PL1]

Has anyone else noticed the deliberate mistake in this 1994 calendar/poster?

--
   -------------------------------------------------------------------------
  |  Richard Herring                |  uucp: <world>!mcvax!ukc!gec-mrc!rnh  |
  |  GEC-Marconi Research Centre    |  Other: rnh@gec-mrc.co.uk             |
  |  GEC-Marconi Ltd, Great Baddow, |  Tel: +44 245 473331 x 3145           |
  |  Chelmsford, Essex, UK CM2 8HN  |  Fax: +44 245 475244 or 478639        |
   -------------------------------------------------------------------------


From epb@avs.com (Ed Bagdonas)
Newsgroups: comp.graphics.avs
Subject: Sales Engineer Positions at AVS (2)
Date: 23 Feb 1994 20:15:45 GMT
Organization: Advanced Visual Systems Inc.
Lines: 36
Message-ID: <2kgdhh$481@nda.nda.com>
NNTP-Posting-Host: gea.avs.com
X-Newsreader: TIN [version 1.2 PL1]


Advanced Visual Systems, of Waltham MA has two open reqs.

District Sales Engineers (2) 
The positions are located in Detroit and Washington, D.C.

Sales Engineers to be responsible for all technical aspects of AVS 
software sales to scientific and engineering customers. 

Each Sales Engineer will team with their respective District 
Sales Representative to develop AVS and UNIRAS prospects and 
customers. 

Task will include:
     -presentation and demonstrations of AVS and UNIRAS.
     -working scientists and engineers of existing install base
     -help determine product requirements for AVS engineering.

Approximately 30% regional travel will be required. 

Job requirements include:
    -Bachelor's of Science and 5+ years of experience as a pre-sales support
     engineer/analyst for workstation/graphics.
    -Must be a strong C/C++ programmer with application development experience.

Please respond by either email: larry@avs.com (ascii only please).
Or via US Post to:   
              AVS Personnel
              300 Fifth Ave.
              Waltham, MA 02154

No phone calls please. Candidates only.  

Advanced Visual Systems is an equal opportunity employer.

 


From ferguson@craycos.com (Scott Ferguson)
Newsgroups: comp.graphics.avs
Subject: Re: Geometry to image
Date: 23 Feb 1994 13:51:07 -0700
Organization: Cray Computer Corporation
Lines: 17
Message-ID: <2kgfjr$on8@nack.craycos.com>
References: <1994Feb21.231019.28830@cis.uab.edu>
NNTP-Posting-Host: nack.craycos.com


>I have modified the geometry of one of my object like changing the color and 
>lighting, and normalize etc. I want to save that particular image. Is there
>any feature under geometry viewer I can save that particular images?

If you invoke the geometry viewer from the Network Editor, you can connect
the output port on the Geometry Viewer module to the input port of the
Image Viewer Module. Then you can save the image from the Image Viewer.
You'll have to update the display once before the image viewer knows there's
something there, since the geometry viewer doesn't produce anything on the
output port unless there's a connection there.

-- 
------------------------------------------------------------------------------
Scott Ferguson                                      Cray Computer Corporation
ferguson@craycos.com                                Colorado Springs, CO 80906
---------My views do not necessarily reflect the views of my employer.--------


From thompson@serts.gsfc.nasa.gov (William Thompson)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Re: problems plotting LARGE amounts of 2D data?
Date: 23 Feb 94 20:20:02 GMT
Organization: NASA Goddard Space Flight Center -- InterNetNews site
Lines: 65
Distribution: world
Message-ID: <thompson.762034802@serts.gsfc.nasa.gov>
References: <2kd84n$aac@daffy.cs.wisc.edu> <thompson.761937768@serts.gsfc.nasa.gov> <2ke290$fac@daffy.cs.wisc.edu>
NNTP-Posting-Host: serts.gsfc.nasa.gov
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

mcheng@dunlop.cs.wisc.edu (Michael Cheng) writes:

>In article <thompson.761937768@serts.gsfc.nasa.gov> thompson@serts.gsfc.nasa.gov (William Thompson) writes:

>>If a program such as IDL (or any of the others you mention) is putting things
>>into virtual memory, it can only be because you don't have enough real memory
>>available to you.  The only thing you can do is buy more memory, or if your
>>operating system supports memory usage quotas (such as VMS) then you need to
>>increase your quotas.

>I argue that buying more memory is not always the best solution for the
>following reasons:
>1) Some of us are poor. (arguably a weak reason)
>2) Some data are always larger than the largest amount of memory 
>reasonable amounts of money can buy.
>3) Even a few medium sized data can overload real memory quickly.
>For example, working with five 25-meg data sets already requires 125 megs
>of memory.

>There has always been a mistmatch between virtual memory policy and
>the need to handle large amounts of data. This mismatch has been demonstrated
>time and again in database systems. I feel that this issue will come
>up again with respect to scientific data sets.

>So we go back to my original question: is there any software package
>for plotting large amounts of 2D data that does better than loading
>everything into virtual memory?

>Mike

Somebody else mailed me privately that he thought that you were probably
talking about database management techniques, rather than memory management
once data had already been read into arrays.  I don't know how the other
packages you mentioned work, but I can comment on how this applies to IDL.

IDL doesn't incorporate any kind of database management scheme, neither
relational nor network nor object-oriented.  It only reads in what you tell it
to read in.  Generally speaking you have to write an IDL routine to read your
data files, although there is built-in support nowadays for certain kinds of
commonly-used file formats such as FITS, HDF, CDF and netCDF.  So, in that
sense IDL does *not* read everything into memory.  It reads in what you tell
it, whether that's an entire file all at once or piece by piece, because you've
written an old-fashioned program to do just what you wanted to do.  You have
complete control and complete responsibility.  :^)

On the other hand, if you wanted to read in a bunch of data and extract global
properties from the data, then it would be easier to read the data into a
single large array.  For example if you had a 2048x2048 image that you wanted
to display, then you would have to read that in as a single large array.  Or if
you wanted to read in 1000 X,Y traces of 500 points each and calculate the
average trace, then it would in general be quicker to read those traces into a
couple of big arrays than to process each trace individually in a loop.  That's
because loops are rather expensive in IDL.

I've certainly done things in IDL where I've handled large amounts of data by
reading in the data bit by bit.  For example, at one point I had a whole series
of images and I wanted to get the mean and standard deviation as a function of
pixel position.  I didn't have enough memory to read all the data at once, so I
read the images one-by-one and did a running calculation.  Other situations is
where I wanted to apply the same routine to a bunch of files.  I just
constructed a program with a FOR loop that read each file in turn and processed
it separately.  In general, though, it's more efficient to use as much memory
as you can.

Bill Thompson


From afj@DrMemory.nuc.ucla.edu (Andy Jacobson)
Newsgroups: comp.graphics.avs
Subject: EPS or PS files
Date: 23 Feb 1994 23:56:58 GMT
Organization: UCLA Dept. of Pharmacology, Los Angeles, CA 90024-6948
Lines: 15
Sender: afj@DrMemory.nuc.ucla.edu (Andy Jacobson)
Message-ID: <2kgqga$6um@news.mic.ucla.edu>
Reply-To: afj@DrMemory.nuc.ucla.edu
NNTP-Posting-Host: drmemory.nuc.ucla.edu

Hi Folks,
Does anyone know of a good way to read in PostScript images to AVS
(or to any conversion program for that matter)? The 'read_eps' and 
'READ ANY IMGE' modules don't seem to work at all for this purpose.

What I am trying to do is (on a Sparc10, SunOS4.1.3, AVS 5.01) edit some saved
postscript files from the "print Network" button in the network editor.
These postscript files print just fine, and come up ok in Sun's
pageview, but go nowhere else. I can't get any of the Adobe software
on the Mac to touch them (except as a separation, and from there just
to the printr). I would prefer not to have to screen dump them if there
is some better way to handle the postscript directly.
Thanks,
-A.J.



From slidel@bsm.biochemistry.ucl.ac.uk (Timothy Slidel)
Newsgroups: comp.graphics.avs
Subject: Online AVS documentation
Date: 23 Feb 1994 18:44:04 -0600
Organization: Bloomsbury Computing Consortium
Lines: 24
Sender: daemon@cs.utexas.edu
Message-ID: <1994Feb23.142131.42436@ucl.ac.uk>
NNTP-Posting-Host: cs.utexas.edu

Does anyone have an online, ftpable version of the documentation for AVS 5 ?
(preferably in ASCII format) - I'm fed up with manually searching for that
paragraph I read the other day !!

Cheers,

Tim.

-- 
**************************************************
Tim Slidel

Biomolecular Structure and Modelling Unit (BSM),
Dept. Biochemistry and Molecular Biology,
University College London,
Gower Street,
London WC1E 6BT.

Tel: (071) 387 7050 x3896
Fax: (071) 380 7193
Email: slidel@uk.ac.ucl.bioc.bsm
**************************************************




From JHHARDE@erenj.com (John Hardenbergh)
Subject: Re: Geometry to image
Message-ID: <JHHARDE.82.2D6CB0D2@erenj.com>
Lines: 42
Sender: news@erenj.com (ERE News System)
Nntp-Posting-Host: answer.erenj.com
Organization: Exxon Research and Engineering
X-Newsreader: Trumpet for Windows [Version 1.0 Rev Final Beta #10]
References: <1994Feb21.231019.28830@cis.uab.edu> <2kgfjr$on8@nack.craycos.com>
Date: Thu, 24 Feb 1994 13:51:14 GMT

In article <2kgfjr$on8@nack.craycos.com> ferguson@craycos.com (Scott Ferguson) writes:
>From: ferguson@craycos.com (Scott Ferguson)
>Subject: Re: Geometry to image
>Date: 23 Feb 1994 13:51:07 -0700


>>I have modified the geometry of one of my object like changing the color and 
>>lighting, and normalize etc. I want to save that particular image. Is there
>>any feature under geometry viewer I can save that particular images?

>If you invoke the geometry viewer from the Network Editor, you can connect
>the output port on the Geometry Viewer module to the input port of the
>Image Viewer Module. Then you can save the image from the Image Viewer.
>You'll have to update the display once before the image viewer knows there's
>something there, since the geometry viewer doesn't produce anything on the
>output port unless there's a connection there.

There's a problem with this approach that I've have yet to find a way around.  
If you change position, lighting etc for a scene you need to save your work as 
a scene ( under cameras ) in order to bring it all back at a later time.  But 
when you read a scene from the geometry viewer, even if you started it from 
the network manager it opens up a new view.  That view appears to be totally 
seperate from the network so when the network executes the output of 
the geometry viewer is the default/first view that was opened and not the 
scence I've read in.  So far the only possable solution I've found is to start 
up AVS with CLI enabled and then use the postscript command to save scene for 
printing etc.  But the "geom_save_postscript" only works with the software 
renderer and there are small differences from the hardware rendered image that 
I don't like.  

>-- 
>------------------------------------------------------------------------------
>Scott Ferguson                                      Cray Computer Corporation
>ferguson@craycos.com                                Colorado Springs, CO 80906
>---------My views do not necessarily reflect the views of my employer.--------

John H. Hardenbergh                  |           .-------.    H     H
Exxon Research and Engineering       |           |       |     \   /
jhharde@erenj.com                    |           |   O   | ---> C=C 
                                     |           |   I   |     /   \
Disclaimer: These are my views,      |           `-------'    H    H
            these are only my views. | Better Chemistry Thru Computer Science


From tzak@brutus.ct.gmr.com (Thomas Zak CT90)
Newsgroups: comp.graphics.avs
Subject: Re: Geometry to image
Date: 24 Feb 1994 14:08:06 GMT
Organization: Electronic Data Systems
Lines: 19
Distribution: world
Message-ID: <2kicc6$h90@rcsuna.gmr.com>
References: <1994Feb21.231019.28830@cis.uab.edu>
Reply-To: tzak@cmsa.gmr.com
NNTP-Posting-Host: brutus.ct.gmr.com

Yun Wang (wang@cis.uab.edu) wrote:
>I have a question about the gemoetry viewer:
>
>I have modified the geometry of one of my object like changing the color and 
>lighting, and normalize etc. I want to save that particular image. Is there
>any feature under geometry viewer I can save that particular images? Or do
>I have to save it as adding .scr and use Network Editor to read that geometry 
>and output to pixmap(or write image)?
>
If you look at the Geometry Viewer module, you can see an image output port
on the bottom right.  Just connect a write image module to that and you're 
set.  Each time the geometry scene is updated, the new image is sent out that
port.  For better performance, leave that port disconnected until you want to 
write to your file.

Thomas E. Zak
Electronic Data Systems
tzak@cmsa.gmr.com
"To make Windows go faster, throw it harder!"


From ebert@cs.umbc.edu (Dr. David Ebert)
Newsgroups: alt.image.medical,comp.graphics,comp.graphics.algorithms,comp.graphics.animation,comp.graphics.avs,comp.graphics.opengl,comp.graphics.explorer,comp.graphics.data-explorer,comp.graphics.visualization,comp.human-factors
Subject: CFP Volume Visualization '94
Date: 24 Feb 1994 14:56:28 -0500
Organization: U. Maryland Baltimore County Computer Science Dept.
Lines: 80
Message-ID: <2kj0pcINN69e@pavlovs.cs.umbc.edu>
NNTP-Posting-Host: pavlovs.cs.umbc.edu


	       1994 Symposium on Volume Visualization

		 October 17-18, 1994 Washington, DC
				   
			Call for Participation

Following our three successful meetings (the Chapel Hill '89, San Diego '90,
and Boston '92 Workshops on Volume Visualization), this fourth meeting 
will provide the opportunity for demonstrations of new developments in this
evolving area. Scientists from all disciplines involved in the visual
presentation and interpretation of volumetric data are invited 
to both submit and attend this Symposium.

The Symposium is sponsored by ACM-SIGGRAPH and the IEEE Computer Society
Technical Committee on Computer Graphics. This Workshop will take place during
the week of October 17-21, 1994 at the Sheraton Premiere at Tyson Center Hotel
in Washington DC area, in conjunction with the Visualization '94 Conference.

Six copies of original material should be submitted to the program co-chairs
on or before March 31, 1994. Authors from North America are asked
to submit their papers to Arie Kaufman. All others are to submit their
papers to Wolfgang Krueger. Suggested topics include, but are not limited to:

* Volume visualization of unstructured and irregular grids.
* Parallel and distributed volume visualization.
* Hardware and software systems.
* Validation and control of rendering quality. 
* Volume segmentation and analysis. 
* Management, storage, and rendering of large datasets.
* User interfacing to volume visualization systems.
* Acceleration techniques for volume rendering.
* Fusion and visualization of multimodal and multidimensional data.
* Visualization of non-scalar volumetric information.
* Modeling and realistic rendering with volumes.
* Discipline-specific application of volume visualization.

Papers should be limited to 5,000 words and may be accompanied by an NTSC
video (6 copies, please).  The accepted papers will appear in the Symposium
Proceeding that will be published by ACM/SIGGRAPH and will be distributed 
to all SIGGRAPH Member "Plus".

Program Co-chairs:
Arie Kaufman			    Wolfgang Krueger
Computer Science Department	    Dept. of Scientific Visualization, GMD-HLRZ
State University of New York	    P.O. Box 1316, Schloss Birlinghoven
Stony Brook, NY 11794-4400	    D-5205 Sankt Augustin 1  GERMANY 
Telephone: 516-632-8441/8428        Telephone: +49 (2241) 14-2367
Fax:	   516-632-8334		    Fax:       +49 (2241) 14-2040 
Email:     ari@cs.sunysb.edu	    Email:     krueger@viswiz.gmd.de

Symposium Co-chairs:
Roni Yagel			    Holly Rushmeier
Dept. of Computer Science	    Rm. B-146, Bldg. 225
The Ohio State University	    NIST
2036 Neil Av. Columbus, OH 43210    Gaithersburg, MD 20899
Telephone: 614-292-0060	    	    Telephone: 301-975-3918
Fax:       614-292-2911		    Fax:       301-963-9137
Email:     yagel@cis.ohio-state.edu Email:     holly@cam.nist.gov

Program Committee:
Nick England 	- University of North Carolina, Chapel Hill
Pat Hanrahan 	- Princeton University
Marc Levoy 	- Stanford University
Bill Lorensen 	- General Electric Co.
Nelson Max 	- Lawrence Livermore National Labs
Greg Nielson 	- Arizona State University
Sam Uselton 	- CS Corp - NASA Ames
Jane Wilhelms 	- University of California at Santa Cruz

Symposium Committee:
David Ebert 	- University of Maryland, Baltimore County
Todd Elvins 	- San Diego Supercomputer Center
Larry Gelberg 	- AVS

-- 
-- Dr. David S. Ebert, Computer Science Department, University of Maryland, --
--- Baltimore County; 5401 Wilkens Ave., Baltimore, MD USA 21228-5398  -------
------ ebert@cs.umbc.edu or ..!{att,pyramid,killer}!cs.umbc.edu!ebert --------
-------------------------------------------------------------------------------


From lipman@cam.nist.gov (Robert_R._Lipman_x2703)
Newsgroups: comp.graphics.avs
Subject: AVS window locations
Message-ID: <10811@fs3.cam.nist.gov>
Date: 24 Feb 94 19:36:11 GMT
Sender: news@cam.nist.gov
Organization: National Institute of Standards and Technology, Gaithersburg, MD
Lines: 18

I'm running AVS 5.01 on SGI workstations (Indigo2 Extreme, Crimson VGX,
Onyx RE) and I'm having a slight problem with window placement.  When
AVS is started up the main menu window comes up so far up in the upper
left corner that the title bar is hidden.  Also when subsequent, geometry
viewer windows come up, they are also stuck way up in the corner.  I'd
like them to appear so that the entire upper and left border of the
window is visible.  I couldn't find anything in the .avsrc file that 
would affect this.  Is there someplace else this behavior is contolled?

Thanks,

Robert Lipman                   ||  E-mail: lipman@cam.nist.gov
National Institute of Standards and Technology (NIST)
Building 225, Room B146         ||  Voice : (301) 975-2703
Gaithersburg, Maryland  20899   ||  Fax   : (301) 963-9137
                                   
PLOKTA - press lots of keys to abort



From jworley@borris.eece.unm.edu (Jeremy Worley)
Newsgroups: comp.graphics.avs,comp.soft-sys.khoros,comp.lang.idl-pvwave
Subject: Re: problems plotting LARGE amounts of 2D data?
Date: 24 Feb 1994 21:51:56 GMT
Organization: Khoral Research, Inc.
Lines: 84
Distribution: world
Message-ID: <JWORLEY.94Feb24145157@borris.eece.unm.edu>
References: <2kd84n$aac@daffy.cs.wisc.edu>
NNTP-Posting-Host: borris.khoros.unm.edu
In-reply-to: mcheng@dunlop.cs.wisc.edu's message of 22 Feb 1994 15:25:11 GMT


   Hi

  michael> I am trying to find ways to plot LARGE amounts of 2D data,
  michael> and I like to know what is currently the state of the
  michael> art. For the sake of this posting, let's say "large" means
  michael> much more data than fits in real memory. From my own
  michael> experience, using the virtual memory of the workstation to
  michael> store large amounts of data impedes performance due to
  michael> excessive paging.  Here is what I have been able to gather
  michael> so far:

  michael> 2) Khoros 1: loads everything into virtual memory.  Any
  michael> updates from Khoros 2.0?

We are actively trying to solve the large data set problem.  It is a
technically challenging problem because the problems posed by large
data sets present different problems in different applications.  For
example, in an isolated application, the problem is pretty much
limited to how to get around the virtual memory limitation
efficiently.  However, in other environments, such as Cantata (the
visual programming language in Khoros), large data sets also pose
problems because intermediate stages in a processing pipeline can
quickly chew up any temp space that may be available.

The first issue is being addressed by data services.  Data services is
a data abstraction that provides read and write data in many file
formats via an application programmers interface (API).  This API
provides a means of storing and retrieving data in units that are
convenient to your application area.  Data services is responsible for
managing memory by caching only the portions of the data set that you
are processing.  Thus, only a reasonable portion of your data set is
in memory at any given time.  Data processing programs (including
graphical applications) that are distributed with Khoros 2.0 will be
written to data services.  People who use Khoros as a development
platform will be strongly encouraged to write their applications to
data services as well.

The second issue, that of multiple copies is a byproduct of the
intermediate stages of the data flow program.  Each operator in the
data flow program is written to read in an input set of data, perform
some processing on the data, and then write out an output set of data.
So, the problem of multiple copies is not really related to the data
flow program, but rather related to the data flow operators.  Ideally,
if the data could be passed between operators in some serial fashion
(such as streams or sockets), intermediate copies of the data would
not be needed.  Unfortunately, data flow operators often can not be
written to accept serial input and produce serial output, but rather
require the ability to access data non-sequentially (an N-dimensional
FFT is an example).  These operators require the ability to randomly
access the data input and output.  As you have pointed out, this
presents a significant problem when operating on large data sets.

Our approach to addressing this problem is to provide functionality
for automatically buffering streamed data so that it can be accessed
in a non-serial fashion.  By addressing this problem in the
infrastructure, we can guarantee that only a minimal number of
temporary copies are present at any time (typically this is two copies
per data pathway).

There are probably some direct ways of "getting around" this problem.
Forcing everyone to write stream-processing routines is one approach.
However we don't see this as a reasonable solution since many
algorithms don't lend themselves to this type of interaction.  Our
goal is to abstract low-level issues away from the users who will be
creating their own modules so that they can focus on the problem of
implementing their algorithms without having to worry about working
around limitations in their hardware and operating system.  In the
context of large data sets and data flow environments such as Cantata,
the cost of doing this is increased overhead in terms of both
temporary storage and performance.  The objective is to minimize these
costs while also minimizing the complexity of the system.

Hope we have been helpful.

		Jeremy Worley, Steve Kubica, and the Khoros Group

	~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	Jeremy Worley			jworley@khoros.unm.edu 
	The Khoros Group		(505)837-6500
--
	~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	Jeremy Worley			jworley@khoros.unm.edu 
	The Khoros Group		(505)837-6500


From JHHARDE@erenj.com (John Hardenbergh)
Subject: Re: Geometry to image
Message-ID: <JHHARDE.83.2D6D199B@erenj.com>
Lines: 40
Sender: news@erenj.com (ERE News System)
Nntp-Posting-Host: answer.erenj.com
Organization: Exxon Research and Engineering
X-Newsreader: Trumpet for Windows [Version 1.0 Rev Final Beta #10]
References: <1994Feb21.231019.28830@cis.uab.edu> <2kicc6$h90@rcsuna.gmr.com>
Date: Thu, 24 Feb 1994 21:18:19 GMT

In article <2kicc6$h90@rcsuna.gmr.com> tzak@brutus.ct.gmr.com (Thomas Zak CT90) writes:
>From: tzak@brutus.ct.gmr.com (Thomas Zak CT90)
>Subject: Re: Geometry to image
>Date: 24 Feb 1994 14:08:06 GMT

>Yun Wang (wang@cis.uab.edu) wrote:
>>I have a question about the gemoetry viewer:
>>
>>I have modified the geometry of one of my object like changing the color and 
>>lighting, and normalize etc. I want to save that particular image. Is there
>>any feature under geometry viewer I can save that particular images? Or do
>>I have to save it as adding .scr and use Network Editor to read that geometry 
>>and output to pixmap(or write image)?
>>
>If you look at the Geometry Viewer module, you can see an image output port
>on the bottom right.  Just connect a write image module to that and you're 
>set.  Each time the geometry scene is updated, the new image is sent out that
>port.  For better performance, leave that port disconnected until you want to 
>write to your file.


Thanks, but you missed the point.  Set up simple network with the Geometry 
viewer and the Geom to postscript module and connect the ports as you 
suggested.  You should see the geometry viewer control panel, the network 
workspace and an image viewer (which will be empty).  Now read in a scene you 
saved from a previous AVS session.  Whala you get a new viewer window that has 
nothing to do with the network.  Fire off the network and you will get a 
nice postscript file that contains a blank image.

>Thomas E. Zak
>Electronic Data Systems
>tzak@cmsa.gmr.com
>"To make Windows go faster, throw it harder!"

John H. Hardenbergh                  |           .-------.    H     H
Exxon Research and Engineering       |           |       |     \   /
jhharde@erenj.com                    |           |   O   | ---> C=C 
                                     |           |   I   |     /   \
Disclaimer: These are my views,      |           `-------'    H    H
            these are only my views. | Better Chemistry Thru Computer Science


From ravi@ryutai.co.jp (Ravikiran Aranke)
Newsgroups: comp.graphics.avs
Subject: Re: UCD module
Date: 25 Feb 1994 10:56:07 +0900
Organization: Ryutai Consultants Co., Tokyo
Lines: 11
Distribution: world
Message-ID: <2kjlrn$q82@hera.ryutai.co.jp>
References: <2kdg5k$jou@mane.cgrg.ohio-state.edu>
NNTP-Posting-Host: localhost.ryutai.co.jp

In article <2kdg5k$jou@mane.cgrg.ohio-state.edu> tim@osc.EDU (Tim Rozmajzl) writes:
>However, ucd_contour
>fails when it receives the UCD structure from my module.  (fatal

Try "ucd reverse cell" module. Perticulary, the repair topology option.
The documentation mentions simiilar problem.
(Haven't tried it myself; instead fixed the node ordering to match
that of AVS.)
-- 
-----------------------------------------------------------------
ravi@ryutai.co.jp


From harm@tessella.co.uk (Mark Harrison)
Subject: Re: Geometry to image
Message-ID: <CLrvwD.GnG@tessella.co.uk>
Organization: Tessella Support Services plc, Abingdon, England
References: <1994Feb21.231019.28830@cis.uab.edu> <2kicc6$h90@rcsuna.gmr.com> <JHHARDE.83.2D6D199B@erenj.com>
Date: Fri, 25 Feb 1994 09:05:01 GMT
Lines: 19

>..stuff deleted...................  Whala you get a new viewer window that has 
>nothing to do with the network.  Fire off the network and you will get a 
>nice postscript file that contains a blank image.
>

You can edit the scene file created by the Geometry save scene.  It's just CLI
commands to load geometry and set lights etc.

The first executable line of this is always a geom_create_scene command
which is why you get a new window.  This is needed if you are running the
geom viewer in standalone, but has already been done when you dragged the
g.v. into the network.  Put a # or remove the line totally and you get the
required result.

Hop this helps

Mark




From dburrast@bt-sys.bt.co.uk (Martin Hanley)
Newsgroups: comp.graphics.avs
Subject: Using a SpaceBall with AVS????
Date: 25 Feb 1994 11:36:15 GMT
Organization: British Telecom, Systems Research
Lines: 26
Message-ID: <2kknrf$neb@xenon.bt-sys.bt.co.uk>
NNTP-Posting-Host: sulphur.bt-sys.bt.co.uk
X-Newsreader: TIN [version 1.2 PL2]



Hi,

Has anyone out there interfaced a SpaceBall with AVS?
We have managed to get our spaceball working but can only move the object
with it. Is it possible in AVS ( or with an add on ) to move the users viewpoint
instead of the object. This would enable the user to fly around their data.
(As opposed to their data flying around them!).

Please reply to my e-mail address if possible as I don't always get on to this
news group.

Cheers in antip,

Dave ..........




--
Dave Burraston                                 dburrast@bt-sys.bt.co.uk
Systems Research Division                     
Visualization Research 
British Telecom Research Labs, Martlesham Heath, Ipswich, Suffolk, IP5 7RE, UK                               



From mccau002@acpub.duke.edu (Gary McCauley)
Newsgroups: comp.graphics.avs
Subject: How to convert polygonal database to a tiled surface object?
Date: 25 Feb 1994 16:32:11 GMT
Organization: Duke University; Durham, NC; USA
Lines: 21
Message-ID: <2kl96b$m87@news.duke.edu>
NNTP-Posting-Host: north8.acpub.duke.edu



Very often 3d reconstructions are based on databases of parallel
sections which contain cross-sectional profiles. The profiles
are represented as polygons (lists of xyz's) and the reconstruction
algorithm creates a surface from this database by forming tiles
between planes (sections). 

Surely someone has written a module to create a tiled geom object
from an input file comprised of polygons described in terms of xyz's.

Yes?

Any input would be appreciated.

gary mccauley

Please respond here or e-mail to:  


	mccauley@watt.radonc.duke.edu


From mfrench@olin.es.com (Mike French)
Newsgroups: comp.graphics.avs
Subject: Re: How to convert polygonal database to a tiled surface object?
Date: 25 Feb 1994 21:13:54 GMT
Organization: Evans & Sutherland, Design Systems Division, SLC, UT
Lines: 46
Message-ID: <2klpmi$495@orca.es.com>
References: <2kl96b$m87@news.duke.edu>
NNTP-Posting-Host: 130.187.85.45

In article 718 mccauley@watt.radonc.duke.edu (Gary McCauley) writes:

>
>Very often 3d reconstructions are based on databases of parallel
>sections which contain cross-sectional profiles. The profiles
>are represented as polygons (lists of xyz's) and the reconstruction
>algorithm creates a surface from this database by forming tiles
>between planes (sections). 
>
>Surely someone has written a module to create a tiled geom object
>from an input file comprised of polygons described in terms of xyz's.
>

In general this is a very difficult problem.
I don't know of any AVS module or public domain code that does this.
The main application is in medical imaging, where contours are drawn on 
planar slices obtained from a scanner, then tissues are reconstructed 
from the contour polygons on each slice.
Some references:

Meyers, Shelley and Sloan, "Surfaces from Contours",
ACM Transactions on Graphics, Vol. 11 No. 3 July 1992.

Muller and Klingert, "Surface Interpolation from Cross Sections",
in 
"Focus on Scientific Visualization", 
Hagen, Muller and Nielson (Eds), Springer-Verlag, 1993,
ISBN 0-387-54940-4 (U.S.)
     3-540-54940-4
     ^^^^^^^^^^^^^   seems the 'I'SBN isn't very 'International' :-) 

I can highly recommend the Muller and Klingert text, about 45 pages of 
good algorithmic details, illustrations and summary of an implementation.

Of course, if you just want to reconstruct extruded sections from a CAD
package, with a single polygon per slice, and constant number of
vertices per polygon, then the problem is considerably easier. 
Even if you've lost the vertex mappings between slices you can
assume a least-rotation solution and correlate minimum distances between 
vertices in polygons on neighbouring slices. 
The task gets worse when you allow varying number of vertices per
polygon and worse still if you have to follow branching shapes between
multiple polygons.

Good luck,
Mike


From schiano@indigo.acs.uci.edu (Allen V. Schiano)
Newsgroups: comp.graphics.avs
Subject: Graph Viewer - can you stop it from plotting until you decode the
Date: 28 Feb 1994 19:54:46 GMT
Organization: University of California, Irvine
Lines: 18
Distribution: world
Message-ID: <2kti66$5jd@news.service.uci.edu>
NNTP-Posting-Host: indigo.acs.uci.edu
cc: temilener@uci.edu

columns
Keywords: 

Hi!

I've been trying to get some things to work in the Graph Viewer with no
results.  Here goes:

If I read a set of data from an ASC file (or an inputted 2D field as well)
and set the two columns to be viewed as x and y I have trouble adding a
second graph with say two other columns because it immediately plots the
first part of the second set!  Having to erase the second erroneous graph is
not a solution!

Anyway, is there a way to halt the graph viewer execution until a set time,
kinda like 'disable flow executive' does in the network editor?

Allen Schiano


