krj@na.toronto.edu (Ken Jackson) (07/10/89)
From: nagmag%ukc.ac.uk@NSFNET-RELAY.AC.UK
Date: Wed, 28 Jun 89 20:03:06 EDT
Subject: NAgMAg v89 #16
NAgMAg Wednesday, June 28 1989 Volume 89 Issue 16
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% The official electronic digest of the NAG Users Association %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Today's Topics
Sparse Matrix Package
Multiple Precision Arithmetic
Sparse Matrix Routines
Replies to issue 15
Powers of Sparse Matrices
%%
%% OK so I can't count -- 14,14,16 -- it's just rounding error.
%% At least it generated some traffic!! The numbering of digest
%% has now been `automated' so it should be wrong more often now.
%%
%% Congratulations to Brian Ford on his OBE.
%%
%% I'm almost turning into a pumpkin (midnight) so this digest goes
%% as it is -- watch out lousy mailers!!
%%
%% Tim
---------------------------
Date: Thu, 22 Jun 1989 10:12:20 ITA
From: Valeria Frontini <AFT0@icineca.earn>
Subject: Sparse Matrix Package
In response to Juergen Bruchhaus:
Y12M, by Prof. Zahary Zlatev, is available from the author, and is
faster (in the runs I tried) than the "corresponding" NAG routines
F01BRF/F04AXF.
The package is fully described in the book:
Y12M - Solution of Large and Sparse Systems of Linear Algebraic
Equations
written by Zahari Zlatev, Jerzy Wasniewski, Kield Schaumburg,
edited by G.Goos, J.Hartmanis, Lecture Notes in Computer Science,
Springer-Verlag.
The code is written in standard Fortran and is machine independent.
The actual e-mail address of Prof. Zlatev is:
zlatev at uicsrd.csrd.uiuc.edu
Valeria Frontini
%% See comments after article 3. Tim
---------------------------
Date: 22 Jun 1989 09:46:28-BST
From: rs <sb365@city.ac.uk>
Subject: Multiple Precision Arithmetic
I don't think that this problem can be solved by NAg directly but I hope
that somebody out there might know where I can start looking...
The problem is to add, multiply and modulus very large integers (at least
100 decimal digits). What I think I need is a set of C routines which I can
use in my own programs. I have written add, subtract and multiply using C
but am going bald fast trying to make a modulus function work.
Thanks
Richard Steele
City University Computer Unit
(Janet: rs@uk.ac.city)
%% I have an oldish copy of Brent's mp (multiple precision arithmetic)
%% package (all in Fortran) this will certainly copes with
%% easily with 100 digit numbers. It's rather big but contains an
%% enormous number of multiple precision functions (e.g., tan, tanh and
%% more exotic flavours). The bits you want could probably be stripped
%% out to make a more compact package.
%% There is also a C library call BIGNUM or something like that which
%% has been advertised on the net recently -- I haven't got the details
%% handy (they are probably written on a scrap of paper somewhere).
%% Mail me if you're interested in either. Tim.
---------------------------
Date: Thu, 22 Jun 89 10:15:25 bst
From: Joe Morice <MORICE@prl.philips.co.uk>
Subject: Sparse Matrix Routines
There are some sparse matrix routines available via the NETLIB service; I
think Tim Hopkins can advise about access. There are also various routines
which might be useful in the ACM Collection - the index and information about
how to obtain the codes are published in Transactions on Mathematical
Software.
Joe Morice
%% OK I admit it -- I've got the netlib software database sitting on a
%% disc here. UKC has just got a contract to set up a UK version of the
%% netlib software distribution service -- it should be operational as
%% soon as we can get a big disc in.
%% The Zlatev are in this database as are all the acm codes.
%% If cajoled I might see fit to mail these to Juergen Bruchhaus
%% but I don't want to start a personal s/ware mailing service in
%% advance of netlibuk :-) Tim
---------------------------
Date: Thu, 22 JUN 89 12:59:10 GMT
From: CAROLINE@vax.nag.co.uk
Subject:Replies to issue 15
Replies to issue 15
-------------------
re: Accuracy of F02BJF
------------------
It is difficult to comment on this reliably until we have seen the
program and problem. The routine has been in the library some time
without any reported faults. The disturbing output may be due to a
programming error or ill- conditioning.
re: Sparse Matrix Package
---------------------
NAG is now marketing a Sparse Matrix Libray prepared by Harwell
Laboratory and containing routines selected from the Harwell Subroutine
Library. The library is distributed as source-text, in standard
conforming Fortran 77 and in both single and double precision. A
complete set of example programs and test programs are provided. There
is a one-volume manual of user-documentation, and an Installers' Note.
An annual licence fee is charged and support is provided.
re: Usage Statistics
----------------
Tim is quite correct. All usage statistics have been produced at
sites, by amendments to the distributed NAG Library. Such changes are
very much machine- dependent and can produce different modes of
counting. For example some counts record when a NAG routine is loaded,
others each time it is executed. (The latter produces large figures
for the G05 routines as might be expected!)
re: Birthday Honours
----------------
We are pleased to report that Brian Ford has been awarded an OBE in the
birthday honours list, alongside such greats as Fred Trueman (*). This
is obviously a great honour for Brian, his family and NAG.
(*) I'm afraid I couldn't quite read David's writing here, and not
being a cricketing fan (despite being born in Yorkshire) I'm not sure
how to spell Fred's name. Apologies to all true Yorkshiremen if it's
wrong. Caroline.
%% It's spelt ok -- as Fred said on the radio recently "With your
%% knowledge of the game you should be on the board of selectors!"
%% Sorry -- couldn't resist that!!
%% Tim
---------------------------
Date: 22-JUN-1989 13:40:29 GMT
From: CMAAE08@vaxa.cc.imperial.ac.uk
Subject:Powers of Sparse Matrices
Re Sparse matrices
We have a user who requires a routine to
raise a sparse matrix to a power.
If there any replies to Juergen Bruchhaus request for sparse
matrix routines, could we be made aware of it.
Thanks.
Sinclair Budd
Imperial College Computer Center
University of London
Exibition Road
London SW7 2BX
%% What sort of power? Integer? Fractional? Can we have some examples
%% of size both of matrix and power? Tim.
---------------------------
%% 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.
%%
%% Replies or submissions to nagmag@uk.ac.ukc
%% Distribution changes to nagmag-request@uk.ac.ukc
%%
%% END OF ISSUE
---------------------------
Reposted by
--
Kenneth R. Jackson, krj@na.toronto.edu (on Internet, CSNet,
Computer Science Dept., ARPAnet, BITNET)
University of Toronto, krj@na.utoronto.ca (CDNnet and other
Toronto, Canada M5S 1A4 X.400 nets (Europe))
(Phone: 416-978-7075) ...!{uunet,pyramid,watmath,ubc-cs}!utai!krj