krj@na.toronto.edu (Ken Jackson) (08/04/89)
NAgMAg Thursday, August 3 1989 Volume 89 Issue 19
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% The official electronic digest of the NAG Users Association %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Today's Topics
Numeric-symbolic interface to NAG
Plotting text vertically
NAGUA Conference programme
Request for back issues
A problem with random numbers
Query about D03EEF
%% Today's guest editor is David Morse (drm@uk.ac.ukc). Tim was going to
%% send this issue out before he went on holiday, but the computer was down
%% with a memory fault so he couldn't. Therefore he had to leave sending
%% it in my incapable hands -- no problems yet, all the magic runes have
%% worked! He will be back on 16th August, and will probably send out the
%% next digest between then and the NAGUA conference.
---------------------------
Date: Fri, 21 Jul 89 08:52 EET
From: "HEIKKI APIOLA, PUH 4572731" <APIOLA@opmvax.csc.fi>
Subject: Numeric-symbolic interface to NAG
It was interesting to read David Foulser's notice about CLAM.
I have been working for quite some time on a project which has
similar goals, which I might briefly describe as a combination of a
high level vector and matrix language, symbol manipulation and both
numerical and graphical subroutine libraries.
I have chosen some of the best representatives of each type as the
main building blocks of the environment but the design is open
in the sense that none of them is unreplaceable.
I have taken two high level systems as the "host systems" which
either generate code for calling NAG or other library routines or
execute the generated codes in a completely transparent manner. My
choices for the symbol manipulation system and the matrix-vector
language are Macsyma and APL respectively.
The hardware environment is also as open as possible including
PC-workstation-mainframe-supercomputer-combination. Most of the
development has been done on a VAX-8650 in the Centre for Scientific
Computing (CSC) but at the moment the whole Finnish supercomputing
environment with UNIX workstations,CRAY X-MP,IBM3090+VF, etc.
is under active consideration.
I will give more details in my presentation in the NAGUA-meeting in
September.
David, I am of course interested in more details about CLAM also in the
sense of possibly aquiring it as one of the software tools for the users
of the CSC-facilities.
Heikki Apiola
Centre fo Scientific Computing
Finnish State Computer Centre
PO-BOX 40
SF-02101 Espoo
Finland
e-mail: apiola@csc.fi
tel: +358-0-4572731
---------------------------
Date: Mon, 24 JUL 89 15:15:46 GMT
From: CAROLINE@vax.nag.co.uk
Subject:
Reply to issue 18
Reply to issue 18
-----------------
Re : Plotting text, vertically and otherwise
---------------------------------------
Following my explanation of how to plot text vertically (Nag Mag 18) there
was a request by David Morse, for software that would plot labels for a
vertical axis in a more 'professional' way, with the text running at
90 degrees to the horizontal. The short answer to that is that such
software will not be provided at the new Mark 3 Graphics Library; The
software has already been frozen, and as such no further changes can be
made at this stage.
Having said that, one of the new Mark 3 routines, for contouring of
scattered data, has the option of plotting labels along the contour line,
and text will therefore be plotted in arbitrary text direction (if the
plotting package and graphics device you are using has the capability to
support this). We will look into the possibility of using the underlying
mechanisms to draw vertical text labels; we will let you know the outcome
in a future edition of the Nag Mag.
As for David's remarks about the clearness of documentation, Mark 3 will be
accompanied by a completely revised User Handbook. We have made changes to
many routine documents, in an effort to increase the clarity of
explanations; diagrams have been included to aid in the understanding of
the functionality of routines. (routine J06YHF will have a diagram!) Also
included is a new section on how to use the interface routines to tailor
graphical output to your specific needs. Again, one of the subjects in this
section is how to plot text, using J06YHF or J06ZAF. We believe that the
new Handbook will help you to make more effective use of the Graphics
Library!
Astrid van Maanen
24 July 1989
---------------------------
Date: Tue, 25 JUL 89 10:52:36 GMT
From: CAROLINE@vax.nag.co.uk
Subject:
Replaces earlier item - programme updated again
NAGUA'89 - Conference latest
--------
Just to tempt those of you who have not yet booked, below is the conference
programme in full. The closing date for bookings is 18 August, but
where bookings have already been recieved, confirmations will be sent out from
the beginning of August.
Tuesday 5 September
Tutorial "Human Computer Interfaces"
Professor Ernest Edmonds, LUTCHI Group, University of Loughborough
Wednesday 6 September
9.00 Introductory Remarks
Dr. M. Dowell, Chairman, NAGUA
9.05 NAG Announcements and Future Plans
Dr. B. Ford, Director, NAG Limited, United Kingdom
9.45 The Rise and Fall and Rise of Portable Programs
Dr. S.I. Feldman, Division Manager, Bell Communications Research, USA
10.30 COFFEE
Industrial Applications of NAG Software
11.00 Title to be confirmed
Ms. V. Scadeng, Ryan MacFarland Ltd, London, United Kingdom
11.45 Differential Equation Software in the Petrochemical Industry:
Requirements and Needs
Mr. R.M. Furzeland, Koninklijke/Shell-Laboratorium, Amsterdam,
The Netherlands
12.30 LUNCH
14.00 The NAG Turbo Pascal Library and the NAG Modula-2 Library
Mr. T. Haarmann and Mr. H. Frese, University of Osnabrueck, West Germany
14.30 New Developments for the PC Graphics Interface
Dr. G. Baszenski, University of Bochum, West Germany
15.00 AGAPE: A Screen Management Library for Fortran Programmers
Mr. T. Engels, Solvay, Brussels, Belgium
15.30 TEA
16.00 The NAG Mark 14 Fortran Library
Mr. J.J. Du Croz, NAG Limited, United Kingdom
16.45 Library Management in a Network
Mr. J. Gallop, Rutherford Appleton Laboratory, Chilton, United Kingdom
Thursday 7th September
9.00 The SERC/DTI Engineering Applications of Transputers Initiative
Dr. M. Jane, Rutherford Appleton Laboratory, Chilton, United Kingdom
9.45 IBM Vector/Parallel Experience
Dr. P. Mayes, NAG Limited, United Kingdom
10.30 COFFEE
11.00 Solving Problems with Transputers : Background and Experience
Dr. D. Prior, University of Edinburgh, United Kingdom
11.30 NAG Software at the Danish Supercomputer Centre, UNI*C
Mr. J. Wasniewski, UNI*C, Lyngby, Denmark
12.00 Signal Processing
Dr. J. McWhirter, RSRE, Malvern, United Kingdom
12.30 LUNCH
14.00 APL-MACSYMA-NAG Interface
Dr. H. Apiola, Finnish State Computer Centre, Espoo
14.30 GLIMPSE
Dr. C. O'Brien, NAG Limited, United Kingdom
15.00 The NAG HELP System
Dr. M. O'Donohoe, University of Cambridge, United Kingdom
15.30 TEA
16.00 Software Product Demonstrations
Friday 8 September
9.00 Fortran 8X
Dr. J. Reid, UKAEA Harwell, United Kingdom
9.40 The Mark 3 Graphics Library
Dr. A.E.P.M. van Maanen, NAG Limited, United Kingdom
10.00 Accessing NAG Library routines from the IBM Expert System Shell ESE
Dr. B. Sueselbeck, Westfaelische Willhelms University, West Germany
10.30 COFFEE
11.00 NAGUA Business Session
11.20 User Forum
Mr. N. Mooljee, University of Edinburgh, United Kingdom
12.20 Closing Remarks
Dr. M. Dowell
12.30 LUNCH
---------------------------
Date: Fri, 28 Jul 89 10:22:27 BST
From: A.N.Other@ThisVAX.ThatVAX.TheOtherVAX.ac.uk
Subject: NAGMAG NO 15
I don't seem to have received NAgMAg no X.
Could you please send me a copy?
Thanks,
Dr A.N. Other
%% Requests such as these probably ought to go to nagmag-request@ukc.ac.uk,
%% rather than to the digest (nagmag@ukc.ac.uk) -- it makes it easier
%% to comply with the request, and these messages don't then have to be
%% filtered out of the digest. (David)
---------------------------
Date: Mon, 31 Jul 89 14:20 EET
From: Hannu Karttunen TLP 90-457 2709 <HKARTTUNEN@opmvax.csc.fi>
Subject: A problem with random numbers
Dear Sirs,
I'm writing about a problem with Nag's (version 12) random numbers.
One of our customers complained about a program running on our Cray
X-MP EA/14 se; the program seemed to run correctly except early in the
morning. The traceback indicates that the Nag-routine g05ccf might be
the reason for the problem. To study this rather curious behaviour I
made a small test program, compiled and linked it with the following
UNICOS commands:
cft77 -es test.f
segldr -o test -l nag test.o
The following compiler listing shows the program and the defaults
of the compiler options:
1TEST PAGE 1 Cray FORTRAN CFT77 3.0X265
04/18/89 13:40:59 07/20/89 08:35:50 PAGE 1 TEST
( -C CRAY-XEA,noema,cigs,vpop -e Bpqrs -d ADISacfghjmoxz
-o full,nozeroinc -i 46 -a stack -t 00 )
1 1. program test
2 2. call g05ccf
3 3. s=0.0
4 4. do 1 i=1,1000
5 5. s=s+g05caf(x)
6 6. 1 continue
7 7. print *,s
8 8. end
I made a crontab entry to run the program every half an hour, each
time appending its output to a log file. It worked well till midnight,
but generated floating point exceptions between 0:00 and 9:30. The
following list shows the critical parts of the log:
...
----------------------------
Wed Jul 19 23:30:00 EET 1989
479.1981447632
----------------------------
Thu Jul 20 00:00:01 EET 1989
TB001 - BEGINNING OF TRACEBACK
- $TRBK WAS CALLED BY f$sig AT 2257d (LINE NUMBER 280)
- f$sig WAS CALLED BY CALLFUNC AT 50002a
- CALLFUNC WAS CALLED BY G05CAF AT 132b
- G05CAF WAS CALLED BY G05CCF AT 262c (LINE NUMBER 150)
- G05CCF WAS CALLED BY TEST AT 57c (LINE NUMBER 2)
- TEST WAS CALLED BY $START$ AT 45a
TB002 - END OF TRACEBACK
----------------------------
...
----------------------------
Thu Jul 20 09:30:01 EET 1989
TB001 - BEGINNING OF TRACEBACK
- $TRBK WAS CALLED BY f$sig AT 2257d (LINE NUMBER 280)
- f$sig WAS CALLED BY CALLFUNC AT 50002a
- CALLFUNC WAS CALLED BY G05CAF AT 132b
- G05CAF WAS CALLED BY G05CCF AT 262c (LINE NUMBER 150)
- G05CCF WAS CALLED BY TEST AT 57c (LINE NUMBER 2)
- TEST WAS CALLED BY $START$ AT 45a
TB002 - END OF TRACEBACK
----------------------------
Thu Jul 20 10:00:02 EET 1989
492.8691542171
Obviously the random number initializer uses the wall-clock time and
fails when the first digit is zero.
Regards
Hannu Karttunen
Centre for Scientific Computing
Espoo, Finland
---------------------------
Date: Mon, 31 JUL 89 15:09:44 GMT
From: CAROLINE@vax.nag.co.uk
Subject:
Reply to issue 17
Further reply to issue 17
-------------------------
re: Queries about D03EEF
--------------------
We would like to thank Professor Proskurowski for his comments on the
NAG multigrid routine D03EEF and its documentation.
Firstly a few words on the philosophy behind the example programs to be
found in the NAG Library manual. These programs are really intended to
give a simple example of how to set up a problem, call the routine and
interpret the results. They are not designed to show the full
capabilities of any particular routine, nor are they intended to show
the behaviour of the routine on particularly "difficult" or "special"
examples. In addition, they are not intended to test the implementation
of the underlying numerical algorithm. For this we have a separate set
of "stringent" test programs.
Turning to the particular routine D03EEF, the example chosen has a
quadratic solution, with zero truncation error for central differences.
This has the unintentional advantage that it gives a demonstration of
the correctness of the implementation in this case. In fact, the results
given by Professor Proskurowski are in good agreement with the multigrid
theory. However, the example has the disadvantage that it gives the
wrong impression about what the user is likely to see from his/her own
problem. It would have been better to give as an example a problem for
which second-order differences gave only second-order accuracy.
We are currently in the process of finalising code and documentation for
Mark 14, and we will change the example to treat a less "special" case.
Peter Mayes
---------------------------
%% For further information about the NAG Users Association please contact:
%% Janet Bentley, Administrator NAGUA,
%% Shore Lane Farm, Blackstone Edge Old Road,
%% LITTLEBOROUGH, Lancashire, OL15 0LQ, UK.
%%
%% or
%% Caroline J Foers
%% NAGUA and Service Coordinator, NAG Ltd,
%% Wilkinson House, Jordan Hill Road,
%% Oxford, OX2 8DR, U.K.
%%
%% e-mail caroline@uk.co.nag.vax
%%
%% Replies or submissions to nagmag@uk.ac.ukc
%% Distribution changes to nagmag-request@uk.ac.ukc
%%
%% END OF ISSUE
-------------------------------
Reposted by
--
Prof. Kenneth R. Jackson, krj@na.toronto.edu (on Internet, CSNet,
Computer Science Dept., ARPAnet, BITNET)
University of Toronto, krj@na.utoronto.ca (on CDNnet and other
Toronto, Ontario, X.400 nets (Europe))
Canada M5S 1A4 ...!{uunet,pyramid,watmath,ubc-cs}!utai!krj