AVL on Mac OS X

NOTE: THIS IS UNSUPPORTED. I ONLY AM POSTING THIS AS A STARTING POINT FOR COMPILING AVL ON YOUR OWN COMPUTER. PLEASE DO NOT SEND EMAILS ASKING FOR ASSISTANCE AS I AM UNABLE TO ASSIST YOU. However, if you have any updates that you would like to contribute, I'd be happy to include them on this page. Thanks - RW 03/13/2010

Here's how I got Athena Vortex Lattice to work on Mac OS X, Leopard. I wanted to play around with it to see how well it'd work for the aircraft design class I have at Illinois Institute of Technology. However, the current version doesn't have any suggestions for building in Mac OS X, so I'd like to give many many props to my good friend, Joshua Nedrud (bioneddy), for basically doing all of this. I just get to profit from his work, and now you can too! The directions aren't the prettiest, in fact, I'd use diff for something a little more elegant, but this is better than the non-documentation that currently exists.

Last updated June 10th, 2010 Please scroll to the bottom for details!

1) Launch X11.
This should automagickally launch xterm. If it doesn't, go to the menu bar: Applications > Terminal (or CMD-N). Use this to execute everything -- mainly because in the end, you'll need to have X11 running to fully use AVL.

2) Download AVL and unpack the tarball.
tar -xvzf avl3.27.tar.gz
cd Avl

3) Build Xplot11
cd plotlib

3.1) Edit config.make using your favorite text editor (like vim).
Change the compiler to FC = gfortran (line 55)

3.2) Make!
make

4) Build eispack
cd ../eispack

4.1) Edit the makefile.
Change the compiler to FC = gfortran

4.2) Make!
make

5) Build AVL
cd ../bin

5.1) Edit the makefile
Change compiler to FC = gfortran (line 26)
Change the pltlib to PLTLIB = -I /usr/X11R6/include -L/usr/X11R6/lib -lXaw -lXext -lXmu -LXt -lX11 (line 28)
Comment out the intel between lines 66-71... make it so that it says:
### Intel Fortran Compiler 8.x
#FC = ifort
#FFLAGS = -O
#PLTLIB = -L/usr/X11R6/lib -lX11
#SECOND = second_ifc.f
#FTNLIB = -Vaxlib
Comment out the double precision stuff (this isn't necessary, but it appears as though gfortran doesn't take the -r8 option, so you may end up with double precision "in name only")
#FFLAGS = -O -r8 (line 74)
#PLTOBJ = ../plotlib/libPltDP.a (line 75)

5.2) Make!
make

6) Run!
./avl

NOTE 10/26/2009: I have received information from a user that is using a newer gfortran (v4.3, from MacResearch.org, I believe) has had issues with compiling. He resolved it by removing the -O option on line 27 of the makefile in Avl/bin so that it now reads FFLAGS=. Hope this helps.

NOTE 06/10/2010: Emmett Lalish of U. of Washington sent in the following: Thanks for the info on compiling AVL for OSX. I ran into a problem and found the solution and thought you might want to add it. I'm using 10.6, which caused an error compiling plotlib (basically: ranlib archive member cputype does not match previous archive member cputype). Looks like this is because 10.6 has gone to 64-bit. Adding a compile flag fixed it right up: just add -m32 to line 61 in config.make so it reads:
CFLAGS = -m32 -O3 $(DEFINE)

Just an aside, I've found there's no need to open X11 first. I just open Terminal (a little nicer than xterm anyway) and run avl; when I ask it to plot something it starts up X11 automatically.

Thanks, -Emmett
No, thank you, Emmett!

enjoy!