Install ip v5.0.0 #9
Comments
|
@DavidHuber-NOAA The ip/5.0.0 has been installed on acorn for preparation testing. Please load and test: |
|
Thank you @Hang-Lei-NOAA. Unfortunately, I do not have a good way to test this installation on Acorn. The GSI, UFS_Utils, and GSI-utils all use this library, but they are not currently ported to Acorn and I'm not sure how extensively these applications would test the library. @edwardhartnett @AlexanderRichert-NOAA Could you point me to a test (tests) to verify that the IP library is working correctly? |
|
@david Our unit test cannot guarantee its working in the model. Therefore,
modeling test is the necessary step for verification.
…On Mon, Sep 16, 2024 at 3:59 PM David Huber ***@***.***> wrote:
Thank you @Hang-Lei-NOAA <https://github.com/Hang-Lei-NOAA>.
Unfortunately, I do not have a good way to test this installation on Acorn.
The GSI, UFS_Utils, and GSI-utils all use this library, but they are not
currently ported to Acorn and I'm not sure how extensively these
applications would test the library.
@edwardhartnett <https://github.com/edwardhartnett> @AlexanderRichert-NOAA
<https://github.com/AlexanderRichert-NOAA> Could you point me to a test
(tests) to verify that the IP library is working correctly?
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFGZGLYKIY4BL4IONL3ZW42ANAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNJTHAYDAMBVG4>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
@DavidHuber-NOAA would you be able to use the stack-stack 1.7.0 or 1.8.0 installations on Acorn to test? Both have ip 5.0. |
|
Alright, I will try getting the GSI on Acorn. Working with the spack-stack installation there should make that effort much easier. |
|
@AlexanderRichert-NOAA I do not see a module file for crtm under |
|
Ah, disregard. I see I should be looking in the mpich suite of modules. |
|
@DavidHuber-NOAA did you get to ip-5.0.0 and did it work for you? |
|
@edwardhartnett I need to finish porting/testing the GSI on Acorn first, then I can test this installation. I hope to finish with all of that by the end of the week. |
|
Opened GSI port to Acorn PR NOAA-EMC/GSI#793. |
|
I have been on sick leave the latter half of the week, so I won’t get to this today. I’ll try again next week. Briefly: initial results show differences between ip 2.3.3 and 5.0 in (IIRC) the RRFS regression test. |
|
@Hang-Lei-NOAA As a prerequisite to porting the GSI to Acorn, I will also need CRTM v2.4.0.1 installed there. This version is already installed on Cactus and Dogwood, so I would only need the Acorn install. Should I open another request for that? |
|
All software installed on cactus must have been installed on acorn already.
No test no delivery.
@david Huber - NOAA Affiliate ***@***.***> please remember that
/lfs/h1/emc/nceplibs/noscrub/hpc-stack/libs/hpc-stack/modulefiles is our
EMC maintained library structure on acorn, It should contain the largest
available sets for EMC use.
Please check out
module
use /lfs/h1/emc/nceplibs/noscrub/hpc-stack/libs/hpc-stack/modulefiles/compiler/intel/19.1.3.304
module load crtm/2.4.0.1
…On Wed, Oct 2, 2024 at 1:59 PM David Huber ***@***.***> wrote:
@Hang-Lei-NOAA <https://github.com/Hang-Lei-NOAA> As a prerequisite to
porting the GSI to Acorn, I will also need CRTM v2.4.0.1 installed there.
This version is already installed on Cactus and Dogwood, so I would only
need the Acorn install. Should I open another request for that?
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFFKWOWPQINF4QWYKODZZQYBHAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOBZGI4TINRZHE>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Ah, very good. Thanks for the reminder @Hang-Lei-NOAA. |
|
@Hang-Lei-NOAA I tested IP v5.0.0 within the GSI and all results are identical to IP v3.3.3/SP v2.3.3, so I think we are good to go. |
|
Okay, I will go ahead to process it.
…On Fri, Oct 4, 2024 at 11:50 AM David Huber ***@***.***> wrote:
@Hang-Lei-NOAA <https://github.com/Hang-Lei-NOAA> I tested IP v5.0.0
within the GSI and all results are identical to IP v3.3.3/SP v2.3.3, so I
think we are good to go.
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFCATFXQAXZZ2ZWRKRLZZ2ZTTAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOJTHE4DQNJRGA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Has this been submitted to NCO? |
|
Not yet, any other issues on this?
…On Mon, Oct 7, 2024 at 9:15 AM Edward Hartnett ***@***.***> wrote:
Has this been submitted to NCO?
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFE6QDPTKK7L4YRN2N3Z2KCOPAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOJWHEYDIMJUGE>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
The R2O delivery instruction is : |
|
Any progress on this issue? |
|
It’s still on NCO’s waiting list for installation. |
|
@DavidHuber-NOAA NCO has installed this version on dogwoods. Please test it: |
|
@Hang-Lei-NOAA Thanks for the info. I am able to load the module, but the ip_ROOT=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0
PWD=/lfs/h2/emc/global/noscrub/David.Huber/GSI/gsi_ip_bufr/ush
IP_INC4=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_4
IP_INC8=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_8
IP_INCd=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_d
IP_LIB8=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_8.a
IP_LIB4=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_4.a
IP_LIBd=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_d.aBut the directory |
|
@david Huber - NOAA Affiliate ***@***.***>
The NCO require us to test based on their suggested method:
===========================
If the variable $HPC_OPT is set to /apps/ops/*prob*/libs, the module will
point you to "/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0". I installed
the libraries in "/apps/ops/para/libs" for your review, so "/apps/ops/*para*
/libs/intel/19.1.3.304/ip/5.0.0" is populated on both machines. In my
environment, the HPC_OPT variable is set by module "envvar/1.0", but yours
may be set elsewhere. Please let me know if using "export
HPC_OPT=/apps/ops/para/libs" corrects the module paths and how
testing goes. I can submit the RFC tomorrow if all goes well.
============================
Could you please try this method?
Or
I have copied their modulefiles and set direct paths
==================my suggested method======================
module use /lfs/h2/emc/eib/save/Hang.Lei/nco
module load ip/5.0.0
module load bufr/12.1.0
====================================
…On Fri, Dec 6, 2024 at 2:52 PM David Huber ***@***.***> wrote:
@Hang-Lei-NOAA <https://github.com/Hang-Lei-NOAA> Thanks for the info. I
am able to load the module, but the ip library does not seem to be
installed in the correct location. The following environmental variables
are set after loading
/apps/ops/para/libs/modulefiles/compiler/intel/19.1.3.304/ip/5.0.0.lua:
ip_ROOT=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0PWD=/lfs/h2/emc/global/noscrub/David.Huber/GSI/gsi_ip_bufr/ushIP_INC4=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_4IP_INC8=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_8IP_INCd=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/include_dIP_LIB8=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_8.aIP_LIB4=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_4.aIP_LIBd=/apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0/lib64/libip_d.a
But the directory /apps/ops/prod/libs/intel/19.1.3.304/ip/5.0.0 does not
exist. Here are the modules loaded at the time of checking this:
1) craype-x86-rome (H) 2) libfabric/1.11.0.0. (H) 3) craype-network-ofi (H) 4) envvar/1.0 5) PrgEnv-intel/8.3.3 6) intel/19.1.3.304 7) craype/2.7.17 8) ip/5.0.0
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFC6AYGBZN6AQ6R2PKD2EH57TAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMRUGA2DKMBWHE>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Thanks for forwarding this info, Hang. First, I tried testing following NCO's suggested method: > module purge
> module reset
> module list
Currently Loaded Modules:
1) craype-x86-rome (H) 2) libfabric/1.11.0.0. (H) 3) craype-network-ofi (H) 4) envvar/1.0
Where:
H: Hidden Module
> echo $HPC_OPT
/apps/ops/prod/libsSo it seems that > module unload envvar
> export HPC_OPT=/apps/ops/para/libs
> module use /apps/ops/para/libs/modulefiles/compiler/intel/19.1.3.304
> module load ip/5.0.0
> module load bufr/12.1.0
> which debufr
/apps/ops/para/libs/intel/19.1.3.304/bufr/12.1.0/bin/debufrIs there another While that is being answered, I will test the GSI against these versions of IP and BUFR. |
|
|
|
Disregard. I forgot that I had removed the sp module as ip/5.0.0 takes on sp functionality. I will continue testing with NCO's modules. |
|
@david Huber - NOAA Affiliate ***@***.***> Thanks, You did a
great job.
…On Mon, Dec 9, 2024 at 11:15 AM David Huber ***@***.***> wrote:
Disregard. I forgot that I had removed the sp module as ip/5.0.0 takes on
sp functionality. I will continue testing with NCO's modules.
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFGCHRXZDIS5E7YGTQD2EW62NAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMRYGUZTKOJZGY>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
@Hang-Lei-NOAA I have finished testing If we can get that library installed to |
|
@david Huber - NOAA Affiliate ***@***.***> Thanks Dave for the
quick test. I will go ahead working with NCO to make sure the prod is good
to go.
…On Fri, Dec 13, 2024 at 9:58 AM David Huber ***@***.***> wrote:
@Hang-Lei-NOAA <https://github.com/Hang-Lei-NOAA> I have finished testing
BUFR 12.1.0 on Dogwood and the library is performing as expected. Thanks!!
If we can get that library installed to /apps/ops/prod/libs or if NCO can
provide a different envvar module, then I think we are good to go here.
—
Reply to this email directly, view it on GitHub
<#9 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKWSMFH2X6UAHIA3OS4JPY32FLYZTAVCNFSM6AAAAABOGAPDOKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNBRGY2DAOJTGE>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Done by NCO |
The new ip library version acquired the all sp functionallity, rendering the latter obsolete. It would be convenient for multiple repositories to have ip v5.0.0 installed on all platforms so that CMake recipes can be updated to exclude sp.
The text was updated successfully, but these errors were encountered: