Re: Installing v5.0 from source

Hi Mary. I went ahead and ran configure and did a "make everything"
per the instructions. After a few false starts, it seems like it ran
fine as ncar graphics version 5 got installed properly, since I see
v5.0 when I run ncargversion. And the "ng4ex gsun01n" test works
like a charm.

But I don't seem to have gotten a new version of ncl. I still have
the previous version of the ncl executable that lives in /usr/local/
bin. Should the installation process have overwritten this one, or
does it put the new one somewhere else that I'm not seeing?

Thanks,
Mark Branson

> Yes it is. If you had gotten anything but a 20, there would be a
> problem. I'm glad to see that Absoft has fixed that problem.
> Hopefully they've fixed they way they do Fortran name mangling
> (especially with regard to Fortran common blocks).
> somk
>
> For anybody else who's interested in the blockdata initialization
> test just for their own curiosity, you can download it from:
>
> http://www.ncl.ucar.edu/Download/files/blockdata_test.tar
>
> You will need to edit the Makefile in the "blockdata_test" directory
> to change the Fortran compiler to the one you are using.
>
> --Mary
>
>
>> Mark
>>
>>> Hi Mark,
>>> Generally you shouldn't need to run "ymake" as a stand-alone
>>> program.
>>> This is done internally when you run "make Everything" or "make me"
>>> or anything else that generates Makefiles.
>>> However, I think what's happening here is that when you run ymake
>>> as a
>>> stand-alone, it needs to know where the configuration files are.
>>> I believe you just need to set the environment variable NCARG_CONFIG
>>> to the directory $NCARG/config, where "$NCARG" represents the top
>>> level of your source tree.
>>> Or, the other thing to try is actually installing the configuration
>>> files. You can do this by typing:
>>> cd $NCARG/config
>>> make install
>>> This may not work, however, depending on how you've customized your
>>> environment. The first method might be your best bet.
>>> Just out of curiosity, have you gotten to the part about the
>>> blockdata
>>> tests? In the past, I've found that the compilers you mention
>>> below do
>>> not handle blockdata initializations properly, and hence you
>>> can't use
>>> them to compile NCL.
>>> --Mary
>>> On Wed, 7 Nov 2007, Mark Branson wrote:
>>>> Hello.
>>>> I'm taking a stab at installing the new v5.0 of ncarg/ncl on my
>>>> Mac PowerPC G5 from the source code. I'm motivated to do this
>>>> because I'm trying to run the CCSM AMWG variability diagnostics
>>>> package and it's failing on the 6-hourly power spectra plots,
>>>> and this part of the package uses fortran programs called by ncl
>>>> routines, so WRAPIT needs to work properly. And we do not use
>>>> GNU fortran on the machines in our research group -- we have
>>>> both the Absoft and XLF compilers available. (From some
>>>> previous posts I've seen on ncl-talk, I suspect that my problems
>>>> are stemming from the fact that we have a g77-compiled version
>>>> of NCL v4.2.0.a032, but I'm using a different fortran compiler
>>>> within WRAPIT to produce the shared object libraries, and that's
>>>> causing some sort of conflict.)
>>>> So I'm following the directions in the INSTALL.html file. I did
>>>> the blockdata test and XLF fails that but Absoft is fine, so
>>>> I'll use the Absoft fortran compiler.
>>>> Now I'm testing ymake and it says:
>>>> ymake : Unable to open NCARG_CONFIG dir : /usr/local/lib/ncarg/
>>>> config
>>>> But I don't think the problem is that it's not recognizing my
>>>> system info, because I can do this:
>>>> /Users/mark/testncl/ncl_ncarg-5.0.0/config> uname -s
>>>> Darwin
>>>> /Users/mark/testncl/ncl_ncarg-5.0.0/config> uname -m
>>>> Power Macintosh
>>>> I should note that this machine currently has ncarg v4.2.0.a023
>>>> and ncl v4.2.0.a032 already installed on it.
>>>> Thanks!
>>>> Mark
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> ncl-talk_at_ucar.edu
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>>
Received on Thu Nov 08 2007 - 14:08:33 MST