[MITgcm-support] build errors for llc_1080

[Ryan Abernathey]

Thanks a lot for the update. Can you just clarify one more thing...

If I run without asyncio, do I just not include that extra code? Is there
anything else that has to be changed?

Thanks,
Ryan

On Wed, Nov 16, 2016 at 10:38 AM, Dimitris Menemenlis <dmenemenlis at gmail.com
> wrote:

> Hi Ryan, I just updated the MITgcm_contrib/llc_hires/llc_*/readme.txt
> files to indicate which is the checkpoint with which the code was last
> compiled and successfully integrated.
>
> llc_1080 simulation is oldest (circa 2013) so compatible with oldest code.
> llc_2160 was last compiled and integrated in 2014
> and llc_4320 is the most current (2016).
>
> If you look at http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/llc_
> hires/llc_4320/code-async/
> there is a record of changes made to bring code_async up to date with
> newer code
>
> code-async is not completely cleaned up yet.  some things need to be
> changed manually.
> see http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/llc_
> hires/llc_4320/code-async/readme.txt
>
> the other thing to keep in mind, as Martin said, is that for llc1080, you
> may not need to use code-async
> usesinglecpuio might be sufficient, depending on machine architecture and
> number of nodes that you are running on.
>
> hope this helps
>
> Dimitris Menemenlis
>
> On Nov 16, 2016, at 4:06 AM, Ryan Abernathey <ryan.abernathey at gmail.com>
> wrote:
>
> Hi,
>
> I'm trying to build the llc_1080 configuration on our local cluster by
> following these instructions from Dimitris:
> http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/llc_
> hires/llc_1080/readme.txt?revision=1.9&view=markup
>
> For my rootdir, I am using the latest fresh checkout of MITgcm from cvs.
> (Maybe this is my mistake.)
>
> I am hitting errors at the build step due to incompatibilities in
> code_async. For example:
> mpiifort -WB -convert big_endian -assume byterecl -O3 -align -march=corei7
> -axAVX  -c asyncio_write_fld_xym_rl.f
> asyncio_write_fld_xym_rl.f(3875): error #6219: This variable, used in a
> specification expression, must be a dummy argument, a COMMON block object,
> or an object accessible through host or use association.   [MULTDIM]
>       Real*8 field(1-Olx:sNx+Olx,1-Oly:sNy+Oly,MULTDIM,nSx,nSy)
> -----------------------------------------------^
> asyncio_write_fld_xym_rl.f(3875): error #6404: This name does not have a
> type, and must have an explicit type.   [MULTDIM]
>
> I fixed this by changing MULTDIM to SEAICE_multDim. But then I hit another
> error
>
> mpiifort -WB -convert big_endian -assume byterecl -O3 -align -march=corei7
> -axAVX  -c write_pickup.f
> write_pickup.f(2334): error #6404: This name does not have a type, and
> must have an explicit type.   [USEDYNP_INEOS_ZC]
>        IF ( useDynP_inEos_Zc ) THEN
> ------------^
>
> I saw that the llc_4320 code_async directory has some more recent
> modifications:
> http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/llc_
> hires/llc_4320/code-async/
> which presumably fix these errors.
>
> Can I safely use  llc_4320/code_async for llc_1080? If not, can you
> suggest how to to update the code to be compatible with recent version of
> MITgcm? Alternatively, maybe you recommend using an earlier checkpoint?
>
> Whatever the case, I thought I would check with the experts before trying
> to plow through on my own.
>
> Thanks,
> Ryan
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20161116/b7f3e16d/attachment.htm>