NASA matrices obtained from Alex Pothen, ICASE and Old Dominion Univ. Converted to HB format by Tim Davis. The following is the original README file, and email discussing the *.nas format. The original *.nas files have coordinate information for the nodes of the graph. They are located in the Conversion/ subdirectory, if you're interested. -------------------------------------------------------------------------------- ftp://ftp.cs.odu.edu/pub/incoming/kumfert/README WARNING: The contents of this ftp site are temporary and subject to change without notice. DISCLAIMER: Nothing free is guaranteed. CONTENTS: The current contents are matrices in various formats. All are compressed and should be ftp'd using binary mode. The two major formats are .nas and .hb for NASA and Harwell-Boeing formats respectively. The current list includes: barth.nas.Z barth4.nas.Z barth5.nas.Z nasa1824.hb.Z nasa2146.hb.Z nasa2910.hb.Z nasa4704.hb.Z nasasrb.hb.Z pwt.nas.Z shuttle_eddy.nas.Z skirt.nas.Z Also included are some files for reading in these problems into the Matlab(R) environment. hbo2mat.f hbo.m readnasa.m Nasa files can be read in using the readnasa.m utility. Loading in Harwell-Boeing files is a two step process. First one must convert the .hb file to a .mat file using hbo2mat. Compilation instructions are included in the file. Then within Matlab, use hbo.m to properly load the .mat file. The latter two files were written by Cleve Moler, The MathWorks. The readnasa.m file was writtem by Gary Kumfert, Old Dominion University. Comments/suggestions can be sent to Gary Kumfert . Matlab is a registered trademark of The MathWorks Inc. -------------------------------------------------------------------------------- Date: Thu, 1 Jun 1995 12:59:00 -0400 From: Pothen Alex To: davis :at the domain: cis.ufl.edu Subject: Re: problems problem.nas is an inhouse format used by us. It lists n coordinates of the nodes x_i y_i (z_i) and then tuples or triples i j a_{ij} The first line lists n (no. of nodes) and nnz (the no. of nonzeros). --Alex From: Alex Pothen Date: Thu, 1 Jun 1995 18:25:15 -0400 To: davis :at the domain: cis.ufl.edu Subject: Re: matrices Tim, I am working at Icase and don't have access to barth I will check and let you know tomorrow. --Alex P.S. If it is a nas file, I trust you are using readnasa.m file to read it. It is an airfoil problem, so you should be able to visualize it using gplot. --Alex From: Alex Pothen Date: Fri, 2 Jun 1995 10:06:46 -0400 To: davis :at the domain: cis.ufl.edu Subject: Re: matrices It is possible that the original numbering of barth makes it look random in a spy plot. --Alex