[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20021102: IMGOLAY/FRMOLAY on Solaris SPARC gcc/g77 does not work (cont.)

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: 20021102: IMGOLAY/FRMOLAY on Solaris SPARC gcc/g77 does not work (cont.)
  • Date: Mon, 04 Nov 2002 15:55:40 -0700
>From: Robert Mullenax <address@hidden>
>Organization: NSBF
>Keywords: 200210270159.g9R1x6q20772 McIDAS v2002 gcc/g77

Robert,

I just found out why FRMOLAY and IMGOLAY fail when built with either
gcc/f2c/mcfc or gcc/g77 on Solaris x86: a buffer dimensioned in
the source code for these routines is larger than can be handled
by these compilation environments.

For the tracking system: the same problem does not occur when using the
Sun compilers.

Here is what you can do to get going:

edit both source files: frmolay.pgm and imgolay.pgm and:

change:

parameter (olaybufsz=10000*9000)

to:

parameter (olaybufsz=10000*7000)       (or anything smaller)

After making the code modifications, you will need to remake and install
the executables for these routines.  For example, when using gcc/g77
you would:

<edit frmolay.pgm and imgolay.pgm>
make frmolay.k imgolay.k VENDOR=-g77
rm ~/bin/frmolay.k ~/bin/imgolay.k
ln frmolay.k imgolay.k ~/bin

I did some searching on Google Groups for this problem and saw that folks
had run into the size limitation in other contexts.  It was my own experiments
that led me to set olaybufsz to be <= 70 10**6.

If these routines still don't work after changing olaybufsize, keep lowering
its value until they do.  The size limitation _may_ be system dependent,
but I am not sure.

Please let me know if the mods work for you.

Tom

>From address@hidden Mon Nov  4 22:12:12 2002

So this would apply to Solaris SPARC as well, right? (my machine)

I will try this Tuesday night on my Ultra 5.

Thanks,
Robert

>From address@hidden Tue Nov  5 08:16:07 2002

On my home machine Sun Ultra 5 (360MHz 256K cache, 256B RAM, Solaris 8)
I just used gcc/f2c (because this was the last thing I tried) and had
to lower the value from 9000 to 5000.  It works..but the color of the
echoes is not correct and it doesn't matter which graphics table I use,
it always comes out the same.  I didn't realize just how slow this
machine was until I did the frmolay..  geez.

I put an image in the incoming ftp directory of the McIDAS account
(FRM.GIF)  I also did this on a Linux box at Universal (RH7.2, built
withh gcc/g77).  It did not require the fix, but again the colors are
wrong.

Thanks,
Robert

>From address@hidden Tue Nov  5 08:17:44 2002

Never mind about the colors..I started McIDAS with 32 graphics levels
and it works fine.

As always, thanks for your help.

Robert